fitcauto
最適化されたハイパーパラメーターをもつ分類モデルの自動選択
構文
説明
fitcauto は、指定した予測子と応答データに基づいて、さまざまなハイパーパラメーターの値をもつ分類モデルのタイプの選択を自動的に試します。既定では、この関数は、ベイズ最適化を使用してモデルとそのハイパーパラメーターの値を選択し、各モデルの交差検証の分類誤差を計算します。最適化が完了すると、fitcauto は、データ セット全体で学習済みの、新しいデータの分類に最適であると予測したモデルを返します。返されたモデルのオブジェクト関数 predict と loss を使用して、新しいデータの分類、およびテスト セットの分類誤差の計算をそれぞれ行うことができます。
データに最適な分類器のタイプがわからない場合は、fitcauto を使用します。分類モデルのハイパーパラメーターを調整するための代替方法については、代替機能を参照してください。
データに含まれる観測値が 10,000 を超える場合は、fitcauto を実行するとき、ベイズ最適化の代わりに非同期連続半減アルゴリズム (ASHA) の使用を検討してください。データ セットの観測値が多いと、多くの場合、ベイズ最適化よりも ASHA 最適化の方が優れた解を速く見つけます。
Mdl = fitcauto(Tbl,ResponseVarName)Mdl を返します。table Tbl には予測子変数と応答変数を格納します。ResponseVarName は応答変数の名前です。
Mdl = fitcauto(___,Name,Value)HyperparameterOptimizationOptions を使用して、ベイズ最適化 (既定) または非同期連続半減アルゴリズム (ASHA) を使用するかどうかを指定します。ASHA 最適化を使用するには、"HyperparameterOptimizationOptions",struct("Optimizer","asha") を指定します。この構造体に追加フィールドを含めて、その他の最適化の側面を制御できます。
[ は、Mdl,OptimizationResults] = fitcauto(___)OptimizationResults も返します。これには、モデル選択とハイパーパラメーター調整のプロセスの結果が含まれます。
例
fitcauto を使用し、table で指定した予測子データと応答データに基づいて、最適化されたハイパーパラメーターをもつ分類モデルを自動的に選択します。
データの読み込み
carbig データ セットを読み込みます。このデータ セットには、1970 年代と 1980 年代初期に製造された自動車の測定値が格納されています。
load carbig米国製かどうかに基づいて、自動車を分類します。
Origin = categorical(cellstr(Origin)); Origin = mergecats(Origin,["France","Japan","Germany", ... "Sweden","Italy","England"],"NotUSA");
Acceleration、Displacement などの予測子変数と応答変数 Origin が格納された table を作成します。
cars = table(Acceleration,Displacement,Horsepower, ...
Model_Year,MPG,Weight,Origin);データの分割
データを学習セットとテスト セットに分割します。モデル選択とハイパーパラメーター調整のプロセスに観測値の約 80% を使用し、fitcauto によって返された最終モデルのパフォーマンスのテストに観測値の 20% を使用します。cvpartition を使用してデータを分割します。
rng("default") % For reproducibility of the data partition c = cvpartition(Origin,"Holdout",0.2); trainingIdx = training(c); % Training set indices carsTrain = cars(trainingIdx,:); testIdx = test(c); % Test set indices carsTest = cars(testIdx,:);
fitcauto の実行
学習データを fitcauto に渡します。既定では、fitcauto は、試行する適切なモデルのタイプを決定し、ベイズ最適化を使用して適切なハイパーパラメーター値を求め、パフォーマンスが最大となることが期待される学習済みモデル Mdl を返します。さらに、fitcauto は、最適化のプロット、および最適化の結果の反復表示を提供します。これらの結果を解釈する方法の詳細については、Verbose の表示を参照してください。
このプロセスにいくらか時間がかかることを見込んでください。Parallel Computing Toolbox™ のライセンスがある場合、最適化を並列で実行するよう指定して最適化プロセスを高速化することを検討してください。これを行うには、名前と値の引数 "HyperparameterOptimizationOptions",struct("UseParallel",true) を fitcauto に渡します。
Mdl = fitcauto(carsTrain,"Origin");Learner types to explore: ensemble, knn, nb, net, svm, tree Total iterations (MaxObjectiveEvaluations): 180 Total time (MaxTime): Inf |=============================================================================================================================================| | Iter | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | result | loss | & validation (sec)| validation loss | validation loss | | | |=============================================================================================================================================| | 1 | Best | 0.10769 | 9.8867 | 0.10769 | 0.10769 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 291 | | | | | | | | | MinLeafSize: 4 | | 2 | Accept | 0.13231 | 8.4208 | 0.10769 | 0.12098 | ensemble | Method: Bag | | | | | | | | | NumLearningCycles: 233 | | | | | | | | | MinLeafSize: 1 | | 3 | Best | 0.10154 | 0.46968 | 0.10154 | 0.10154 | tree | MinLeafSize: 5 | | 4 | Accept | 0.37231 | 0.77391 | 0.10154 | 0.10154 | svm | BoxConstraint: 0.22029 | | | | | | | | | KernelScale: 0.4534 | | 5 | Accept | 0.14769 | 0.12722 | 0.10154 | 0.12098 | tree | MinLeafSize: 2 | | 6 | Accept | 0.24308 | 0.54475 | 0.10154 | 0.12098 | knn | NumNeighbors: 155 | | 7 | Accept | 0.37231 | 0.17065 | 0.10154 | 0.12098 | svm | BoxConstraint: 22.3 | | | | | | | | | KernelScale: 0.018132 | | 8 | Accept | 0.19385 | 0.098037 | 0.10154 | 0.12098 | knn | NumNeighbors: 75 | | 9 | Accept | 0.15692 | 8.2932 | 0.10154 | 0.12098 | net | Activations: tanh | | | | | | | | | Standardize: true | | | | | | | | | Lambda: 0.0010226 | | | | | | | | | LayerSizes: 224 | | 10 | Accept | 0.37231 | 0.20214 | 0.10154 | 0.12098 | net | Activations: tanh | | | | | | | | | Standardize: false | | | | | | | | | Lambda: 302.8 | | | | | | | | | LayerSizes: [ 78 5 ] | | 11 | Accept | 0.22769 | 1.2781 | 0.10154 | 0.12098 | nb | DistributionNames: kernel | | | | | | | | | Width: 214.96 | | | | | | | | | Standardize: false | | 12 | Accept | 0.18154 | 5.8459 | 0.10154 | 0.12323 | ensemble | Method: Bag | | | | | | | | | NumLearningCycles: 232 | | | | | | | | | MinLeafSize: 105 | | 13 | Accept | 0.19692 | 0.090905 | 0.10154 | 0.12323 | knn | NumNeighbors: 13 | | 14 | Accept | 0.20308 | 0.10957 | 0.10154 | 0.12323 | svm | BoxConstraint: 9.1458 | | | | | | | | | KernelScale: 18.65 | | 15 | Accept | 0.14154 | 7.7326 | 0.10154 | 0.12323 | ensemble | Method: Bag | | | | | | | | | NumLearningCycles: 278 | | | | | | | | | MinLeafSize: 2 | | 16 | Accept | 0.11692 | 4.5938 | 0.10154 | 0.12323 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 214 | | | | | | | | | MinLeafSize: 1 | | 17 | Accept | 0.18462 | 0.088159 | 0.10154 | 0.12323 | knn | NumNeighbors: 10 | | 18 | Accept | 0.12308 | 0.077702 | 0.10154 | 0.12484 | tree | MinLeafSize: 10 | | 19 | Accept | 0.26462 | 3.68 | 0.10154 | 0.12484 | net | Activations: tanh | | | | | | | | | Standardize: false | | | | | | | | | Lambda: 0.44352 | | | | | | | | | LayerSizes: [ 48 40 ] | | 20 | Accept | 0.10769 | 4.1582 | 0.10154 | 0.12484 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 206 | | | | | | | | | MinLeafSize: 3 | |=============================================================================================================================================| | Iter | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | result | loss | & validation (sec)| validation loss | validation loss | | | |=============================================================================================================================================| | 21 | Accept | 0.12 | 0.092835 | 0.10154 | 0.12349 | tree | MinLeafSize: 13 | | 22 | Accept | 0.19385 | 0.08931 | 0.10154 | 0.12349 | knn | NumNeighbors: 57 | | 23 | Accept | 0.37231 | 0.29219 | 0.10154 | 0.12349 | net | Activations: relu | | | | | | | | | Standardize: true | | | | | | | | | Lambda: 1.6049 | | | | | | | | | LayerSizes: [ 236 2 197 ] | | 24 | Accept | 0.13538 | 7.2642 | 0.10154 | 0.12349 | ensemble | Method: Bag | | | | | | | | | NumLearningCycles: 272 | | | | | | | | | MinLeafSize: 4 | | 25 | Accept | 0.21231 | 0.12119 | 0.10154 | 0.12349 | svm | BoxConstraint: 3.2832 | | | | | | | | | KernelScale: 19.127 | | 26 | Accept | 0.23692 | 0.088582 | 0.10154 | 0.12349 | knn | NumNeighbors: 154 | | 27 | Accept | 0.18154 | 0.071505 | 0.10154 | 0.12413 | tree | MinLeafSize: 30 | | 28 | Accept | 0.22769 | 0.14963 | 0.10154 | 0.12413 | nb | DistributionNames: normal | | | | | | | | | Width: NaN | | | | | | | | | Standardize: - | | 29 | Accept | 0.37231 | 0.31305 | 0.10154 | 0.12413 | net | Activations: tanh | | | | | | | | | Standardize: false | | | | | | | | | Lambda: 9.7449 | | | | | | | | | LayerSizes: [ 123 213 22 ] | | 30 | Accept | 0.22769 | 0.061021 | 0.10154 | 0.12413 | nb | DistributionNames: normal | | | | | | | | | Width: NaN | | | | | | | | | Standardize: - | | 31 | Accept | 0.20615 | 7.6661 | 0.10154 | 0.12413 | net | Activations: sigmoid | | | | | | | | | Standardize: false | | | | | | | | | Lambda: 9.0476e-06 | | | | | | | | | LayerSizes: [ 1 286 ] | | 32 | Accept | 0.37231 | 0.12123 | 0.10154 | 0.12413 | net | Activations: none | | | | | | | | | Standardize: true | | | | | | | | | Lambda: 169.93 | | | | | | | | | LayerSizes: [ 8 43 ] | | 33 | Accept | 0.16923 | 0.053687 | 0.10154 | 0.12413 | tree | MinLeafSize: 50 | | 34 | Accept | 0.12923 | 0.052139 | 0.10154 | 0.12413 | tree | MinLeafSize: 3 | | 35 | Accept | 0.37538 | 0.16254 | 0.10154 | 0.12413 | svm | BoxConstraint: 31.988 | | | | | | | | | KernelScale: 0.068262 | | 36 | Best | 0.086154 | 5.1664 | 0.086154 | 0.11432 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 259 | | | | | | | | | MinLeafSize: 46 | | 37 | Accept | 0.20308 | 0.093464 | 0.086154 | 0.11432 | knn | NumNeighbors: 7 | | 38 | Accept | 0.19692 | 0.062179 | 0.086154 | 0.11432 | knn | NumNeighbors: 9 | | 39 | Accept | 0.21846 | 0.20379 | 0.086154 | 0.11432 | nb | DistributionNames: kernel | | | | | | | | | Width: 11.43 | | | | | | | | | Standardize: false | | 40 | Accept | 0.22769 | 0.077831 | 0.086154 | 0.11432 | nb | DistributionNames: normal | | | | | | | | | Width: NaN | | | | | | | | | Standardize: - | |=============================================================================================================================================| | Iter | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | result | loss | & validation (sec)| validation loss | validation loss | | | |=============================================================================================================================================| | 41 | Accept | 0.24615 | 0.2016 | 0.086154 | 0.11432 | nb | DistributionNames: kernel | | | | | | | | | Width: 0.95982 | | | | | | | | | Standardize: false | | 42 | Accept | 0.22769 | 0.054205 | 0.086154 | 0.11432 | nb | DistributionNames: normal | | | | | | | | | Width: NaN | | | | | | | | | Standardize: - | | 43 | Accept | 0.16308 | 1.8399 | 0.086154 | 0.11432 | net | Activations: tanh | | | | | | | | | Standardize: true | | | | | | | | | Lambda: 8.2888e-08 | | | | | | | | | LayerSizes: 20 | | 44 | Accept | 0.37231 | 0.10773 | 0.086154 | 0.11432 | svm | BoxConstraint: 0.5593 | | | | | | | | | KernelScale: 0.003667 | | 45 | Accept | 0.11385 | 0.075454 | 0.086154 | 0.11432 | tree | MinLeafSize: 11 | | 46 | Accept | 0.20308 | 0.074108 | 0.086154 | 0.11432 | knn | NumNeighbors: 7 | | 47 | Accept | 0.22769 | 0.061325 | 0.086154 | 0.11432 | nb | DistributionNames: normal | | | | | | | | | Width: NaN | | | | | | | | | Standardize: - | | 48 | Accept | 0.25231 | 5.6001 | 0.086154 | 0.11432 | net | Activations: sigmoid | | | | | | | | | Standardize: false | | | | | | | | | Lambda: 4.2329e-05 | | | | | | | | | LayerSizes: [ 16 21 86 ] | | 49 | Accept | 0.22769 | 0.051748 | 0.086154 | 0.11432 | nb | DistributionNames: normal | | | | | | | | | Width: NaN | | | | | | | | | Standardize: - | | 50 | Accept | 0.37231 | 0.11468 | 0.086154 | 0.11432 | svm | BoxConstraint: 47.547 | | | | | | | | | KernelScale: 0.038904 | | 51 | Accept | 0.16923 | 0.074099 | 0.086154 | 0.11432 | tree | MinLeafSize: 76 | | 52 | Accept | 0.15692 | 1.6882 | 0.086154 | 0.11432 | net | Activations: tanh | | | | | | | | | Standardize: false | | | | | | | | | Lambda: 0.24612 | | | | | | | | | LayerSizes: 6 | | 53 | Accept | 0.10462 | 4.0245 | 0.086154 | 0.10962 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 211 | | | | | | | | | MinLeafSize: 6 | | 54 | Accept | 0.11385 | 0.092927 | 0.086154 | 0.10962 | tree | MinLeafSize: 11 | | 55 | Accept | 0.086154 | 5.1988 | 0.086154 | 0.098594 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 286 | | | | | | | | | MinLeafSize: 47 | | 56 | Accept | 0.19692 | 0.075963 | 0.086154 | 0.098594 | knn | NumNeighbors: 28 | | 57 | Accept | 0.12 | 0.070258 | 0.086154 | 0.098594 | tree | MinLeafSize: 4 | | 58 | Accept | 0.2 | 0.068373 | 0.086154 | 0.098594 | knn | NumNeighbors: 3 | | 59 | Accept | 0.37231 | 0.15296 | 0.086154 | 0.098594 | net | Activations: tanh | | | | | | | | | Standardize: false | | | | | | | | | Lambda: 79.211 | | | | | | | | | LayerSizes: 1 | | 60 | Accept | 0.19692 | 0.061209 | 0.086154 | 0.098594 | knn | NumNeighbors: 13 | |=============================================================================================================================================| | Iter | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | result | loss | & validation (sec)| validation loss | validation loss | | | |=============================================================================================================================================| | 61 | Accept | 0.37231 | 5.3523 | 0.086154 | 0.10784 | ensemble | Method: Bag | | | | | | | | | NumLearningCycles: 234 | | | | | | | | | MinLeafSize: 155 | | 62 | Accept | 0.21846 | 0.076683 | 0.086154 | 0.10784 | knn | NumNeighbors: 2 | | 63 | Accept | 0.32308 | 1.0531 | 0.086154 | 0.10784 | net | Activations: none | | | | | | | | | Standardize: false | | | | | | | | | Lambda: 0.00049454 | | | | | | | | | LayerSizes: [ 3 145 ] | | 64 | Accept | 0.11077 | 5.6836 | 0.086154 | 0.10698 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 300 | | | | | | | | | MinLeafSize: 1 | | 65 | Accept | 0.19692 | 2.5205 | 0.086154 | 0.10698 | net | Activations: tanh | | | | | | | | | Standardize: true | | | | | | | | | Lambda: 1.1014e-05 | | | | | | | | | LayerSizes: [ 26 2 ] | | 66 | Accept | 0.12 | 0.072271 | 0.086154 | 0.10698 | tree | MinLeafSize: 4 | | 67 | Accept | 0.21846 | 0.095838 | 0.086154 | 0.10698 | knn | NumNeighbors: 2 | | 68 | Accept | 0.37231 | 0.091219 | 0.086154 | 0.10698 | svm | BoxConstraint: 0.0061987 | | | | | | | | | KernelScale: 31.312 | | 69 | Accept | 0.37231 | 0.064588 | 0.086154 | 0.10698 | svm | BoxConstraint: 0.070644 | | | | | | | | | KernelScale: 0.22263 | | 70 | Accept | 0.37538 | 0.082788 | 0.086154 | 0.10698 | svm | BoxConstraint: 9.2391 | | | | | | | | | KernelScale: 0.062892 | | 71 | Accept | 0.15692 | 0.061431 | 0.086154 | 0.10698 | svm | BoxConstraint: 32.735 | | | | | | | | | KernelScale: 1.3401 | | 72 | Accept | 0.31077 | 0.05396 | 0.086154 | 0.10698 | svm | BoxConstraint: 0.0407 | | | | | | | | | KernelScale: 3.4118 | | 73 | Accept | 0.37231 | 0.053805 | 0.086154 | 0.10698 | svm | BoxConstraint: 0.0019745 | | | | | | | | | KernelScale: 62.689 | | 74 | Accept | 0.13538 | 5.1971 | 0.086154 | 0.10699 | ensemble | Method: Bag | | | | | | | | | NumLearningCycles: 205 | | | | | | | | | MinLeafSize: 3 | | 75 | Accept | 0.22769 | 0.068918 | 0.086154 | 0.10699 | nb | DistributionNames: normal | | | | | | | | | Width: NaN | | | | | | | | | Standardize: - | | 76 | Accept | 0.37231 | 0.06363 | 0.086154 | 0.10699 | svm | BoxConstraint: 0.42987 | | | | | | | | | KernelScale: 234.78 | | 77 | Accept | 0.25846 | 0.2272 | 0.086154 | 0.10699 | nb | DistributionNames: kernel | | | | | | | | | Width: 2.2007 | | | | | | | | | Standardize: false | | 78 | Accept | 0.13846 | 5.0838 | 0.086154 | 0.10344 | ensemble | Method: Bag | | | | | | | | | NumLearningCycles: 207 | | | | | | | | | MinLeafSize: 4 | | 79 | Accept | 0.14462 | 0.071561 | 0.086154 | 0.10344 | tree | MinLeafSize: 21 | | 80 | Accept | 0.37231 | 0.097395 | 0.086154 | 0.10344 | svm | BoxConstraint: 176.96 | | | | | | | | | KernelScale: 0.022732 | |=============================================================================================================================================| | Iter | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | result | loss | & validation (sec)| validation loss | validation loss | | | |=============================================================================================================================================| | 81 | Accept | 0.19692 | 0.07841 | 0.086154 | 0.10344 | knn | NumNeighbors: 9 | | 82 | Accept | 0.26154 | 0.17034 | 0.086154 | 0.10344 | nb | DistributionNames: kernel | | | | | | | | | Width: 2.808 | | | | | | | | | Standardize: false | | 83 | Accept | 0.37231 | 0.18007 | 0.086154 | 0.10344 | nb | DistributionNames: kernel | | | | | | | | | Width: 615.95 | | | | | | | | | Standardize: true | | 84 | Accept | 0.26154 | 1.8058 | 0.086154 | 0.10344 | net | Activations: relu | | | | | | | | | Standardize: false | | | | | | | | | Lambda: 8.2545e-08 | | | | | | | | | LayerSizes: [ 57 2 7 ] | | 85 | Accept | 0.22769 | 0.063696 | 0.086154 | 0.10344 | nb | DistributionNames: normal | | | | | | | | | Width: NaN | | | | | | | | | Standardize: - | | 86 | Accept | 0.14769 | 6.2743 | 0.086154 | 0.10769 | ensemble | Method: Bag | | | | | | | | | NumLearningCycles: 252 | | | | | | | | | MinLeafSize: 32 | | 87 | Accept | 0.17231 | 0.06519 | 0.086154 | 0.10769 | tree | MinLeafSize: 126 | | 88 | Accept | 0.16923 | 0.043935 | 0.086154 | 0.10769 | tree | MinLeafSize: 93 | | 89 | Accept | 0.22769 | 0.18302 | 0.086154 | 0.10769 | nb | DistributionNames: kernel | | | | | | | | | Width: 0.43629 | | | | | | | | | Standardize: false | | 90 | Accept | 0.26462 | 3.9886 | 0.086154 | 0.10769 | net | Activations: sigmoid | | | | | | | | | Standardize: false | | | | | | | | | Lambda: 7.2789e-07 | | | | | | | | | LayerSizes: 98 | | 91 | Accept | 0.098462 | 4.8539 | 0.086154 | 0.10445 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 259 | | | | | | | | | MinLeafSize: 39 | | 92 | Accept | 0.11077 | 4.9109 | 0.086154 | 0.099362 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 273 | | | | | | | | | MinLeafSize: 70 | | 93 | Accept | 0.13846 | 5.1708 | 0.086154 | 0.10034 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 281 | | | | | | | | | MinLeafSize: 84 | | 94 | Accept | 0.12615 | 4.9036 | 0.086154 | 0.099454 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 244 | | | | | | | | | MinLeafSize: 76 | | 95 | Accept | 0.11077 | 5.5508 | 0.086154 | 0.097618 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 285 | | | | | | | | | MinLeafSize: 71 | | 96 | Accept | 0.14154 | 4.4952 | 0.086154 | 0.095919 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 245 | | | | | | | | | MinLeafSize: 57 | | 97 | Accept | 0.095385 | 4.815 | 0.086154 | 0.094528 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 260 | | | | | | | | | MinLeafSize: 51 | | 98 | Accept | 0.12308 | 4.6592 | 0.086154 | 0.093172 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 260 | | | | | | | | | MinLeafSize: 77 | | 99 | Accept | 0.37231 | 4.8941 | 0.086154 | 0.094437 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 268 | | | | | | | | | MinLeafSize: 149 | | 100 | Accept | 0.13538 | 5.2577 | 0.086154 | 0.096586 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 295 | | | | | | | | | MinLeafSize: 92 | |=============================================================================================================================================| | Iter | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | result | loss | & validation (sec)| validation loss | validation loss | | | |=============================================================================================================================================| | 101 | Accept | 0.092308 | 4.4168 | 0.086154 | 0.092045 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 247 | | | | | | | | | MinLeafSize: 43 | | 102 | Accept | 0.13846 | 4.8009 | 0.086154 | 0.093381 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 265 | | | | | | | | | MinLeafSize: 90 | | 103 | Accept | 0.089231 | 4.393 | 0.086154 | 0.091453 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 245 | | | | | | | | | MinLeafSize: 48 | | 104 | Accept | 0.098462 | 4.7966 | 0.086154 | 0.091605 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 265 | | | | | | | | | MinLeafSize: 52 | | 105 | Accept | 0.086154 | 4.9362 | 0.086154 | 0.090054 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 276 | | | | | | | | | MinLeafSize: 47 | | 106 | Accept | 0.089231 | 4.1084 | 0.086154 | 0.088948 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 224 | | | | | | | | | MinLeafSize: 45 | | 107 | Accept | 0.13538 | 5.0023 | 0.086154 | 0.089673 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 279 | | | | | | | | | MinLeafSize: 97 | | 108 | Accept | 0.13538 | 5.0726 | 0.086154 | 0.089434 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 284 | | | | | | | | | MinLeafSize: 99 | | 109 | Accept | 0.14154 | 3.9115 | 0.086154 | 0.089922 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 214 | | | | | | | | | MinLeafSize: 99 | | 110 | Accept | 0.089231 | 5.324 | 0.086154 | 0.089121 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 296 | | | | | | | | | MinLeafSize: 47 | | 111 | Accept | 0.18154 | 5.2688 | 0.086154 | 0.088882 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 293 | | | | | | | | | MinLeafSize: 124 | | 112 | Accept | 0.13231 | 4.3796 | 0.086154 | 0.088506 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 247 | | | | | | | | | MinLeafSize: 54 | | 113 | Accept | 0.098462 | 3.9631 | 0.086154 | 0.087879 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 219 | | | | | | | | | MinLeafSize: 51 | | 114 | Accept | 0.11077 | 5.0266 | 0.086154 | 0.088112 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 278 | | | | | | | | | MinLeafSize: 28 | | 115 | Accept | 0.098462 | 5.1256 | 0.086154 | 0.089033 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 286 | | | | | | | | | MinLeafSize: 37 | | 116 | Accept | 0.089231 | 5.3373 | 0.086154 | 0.08886 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 298 | | | | | | | | | MinLeafSize: 47 | | 117 | Accept | 0.086154 | 4.9772 | 0.086154 | 0.087554 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 273 | | | | | | | | | MinLeafSize: 47 | | 118 | Accept | 0.095385 | 4.9785 | 0.086154 | 0.087693 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 276 | | | | | | | | | MinLeafSize: 46 | | 119 | Accept | 0.086154 | 5.4771 | 0.086154 | 0.087702 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 283 | | | | | | | | | MinLeafSize: 49 | | 120 | Accept | 0.092308 | 3.7618 | 0.086154 | 0.087509 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 205 | | | | | | | | | MinLeafSize: 47 | |=============================================================================================================================================| | Iter | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | result | loss | & validation (sec)| validation loss | validation loss | | | |=============================================================================================================================================| | 121 | Accept | 0.10462 | 4.2801 | 0.086154 | 0.087866 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 236 | | | | | | | | | MinLeafSize: 52 | | 122 | Accept | 0.37231 | 0.10976 | 0.086154 | 0.087866 | net | Activations: tanh | | | | | | | | | Standardize: true | | | | | | | | | Lambda: 0.72122 | | | | | | | | | LayerSizes: 13 | | 123 | Accept | 0.37231 | 0.10272 | 0.086154 | 0.087866 | net | Activations: tanh | | | | | | | | | Standardize: true | | | | | | | | | Lambda: 0.43917 | | | | | | | | | LayerSizes: 2 | | 124 | Accept | 0.2 | 0.12343 | 0.086154 | 0.087866 | net | Activations: tanh | | | | | | | | | Standardize: true | | | | | | | | | Lambda: 0.050009 | | | | | | | | | LayerSizes: 6 | | 125 | Accept | 0.29231 | 0.12056 | 0.086154 | 0.087866 | net | Activations: tanh | | | | | | | | | Standardize: true | | | | | | | | | Lambda: 0.30341 | | | | | | | | | LayerSizes: 8 | | 126 | Accept | 0.089231 | 4.654 | 0.086154 | 0.087499 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 259 | | | | | | | | | MinLeafSize: 48 | | 127 | Accept | 0.086154 | 5.012 | 0.086154 | 0.086595 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 282 | | | | | | | | | MinLeafSize: 49 | | 128 | Accept | 0.18154 | 4.9285 | 0.086154 | 0.08742 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 282 | | | | | | | | | MinLeafSize: 124 | | 129 | Accept | 0.086154 | 5.1215 | 0.086154 | 0.087394 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 277 | | | | | | | | | MinLeafSize: 49 | | 130 | Accept | 0.086154 | 4.8293 | 0.086154 | 0.086574 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 270 | | | | | | | | | MinLeafSize: 49 | | 131 | Accept | 0.13846 | 4.2622 | 0.086154 | 0.086693 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 239 | | | | | | | | | MinLeafSize: 104 | | 132 | Accept | 0.14154 | 4.4707 | 0.086154 | 0.086428 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 252 | | | | | | | | | MinLeafSize: 104 | | 133 | Accept | 0.37231 | 0.23993 | 0.086154 | 0.086428 | net | Activations: tanh | | | | | | | | | Standardize: true | | | | | | | | | Lambda: 0.15163 | | | | | | | | | LayerSizes: [ 20 93 ] | | 134 | Best | 0.083077 | 4.9394 | 0.083077 | 0.085629 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 275 | | | | | | | | | MinLeafSize: 49 | | 135 | Accept | 0.086154 | 4.7512 | 0.083077 | 0.08537 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 266 | | | | | | | | | MinLeafSize: 49 | | 136 | Accept | 0.089231 | 4.1396 | 0.083077 | 0.085748 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 228 | | | | | | | | | MinLeafSize: 49 | | 137 | Accept | 0.14154 | 4.5844 | 0.083077 | 0.08557 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 263 | | | | | | | | | MinLeafSize: 110 | | 138 | Accept | 0.089231 | 4.1878 | 0.083077 | 0.085505 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 233 | | | | | | | | | MinLeafSize: 49 | | 139 | Accept | 0.15385 | 6.4706 | 0.083077 | 0.085674 | ensemble | Method: Bag | | | | | | | | | NumLearningCycles: 273 | | | | | | | | | MinLeafSize: 49 | | 140 | Accept | 0.17538 | 6.311 | 0.083077 | 0.08568 | ensemble | Method: Bag | | | | | | | | | NumLearningCycles: 271 | | | | | | | | | MinLeafSize: 106 | |=============================================================================================================================================| | Iter | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | result | loss | & validation (sec)| validation loss | validation loss | | | |=============================================================================================================================================| | 141 | Accept | 0.083077 | 5.0683 | 0.083077 | 0.085124 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 275 | | | | | | | | | MinLeafSize: 49 | | 142 | Accept | 0.083077 | 4.9466 | 0.083077 | 0.084346 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 272 | | | | | | | | | MinLeafSize: 49 | | 143 | Accept | 0.18462 | 4.0363 | 0.083077 | 0.084296 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 227 | | | | | | | | | MinLeafSize: 115 | | 144 | Accept | 0.098462 | 4.8689 | 0.083077 | 0.084396 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 271 | | | | | | | | | MinLeafSize: 50 | | 145 | Accept | 0.17231 | 0.11083 | 0.083077 | 0.084396 | svm | BoxConstraint: 7.4019 | | | | | | | | | KernelScale: 0.98931 | | 146 | Accept | 0.21538 | 0.059402 | 0.083077 | 0.084396 | svm | BoxConstraint: 17.066 | | | | | | | | | KernelScale: 0.77684 | | 147 | Accept | 0.086154 | 4.8705 | 0.083077 | 0.084203 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 271 | | | | | | | | | MinLeafSize: 48 | | 148 | Accept | 0.086154 | 4.8211 | 0.083077 | 0.084345 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 270 | | | | | | | | | MinLeafSize: 49 | | 149 | Accept | 0.37231 | 0.063284 | 0.083077 | 0.084345 | svm | BoxConstraint: 41.79 | | | | | | | | | KernelScale: 281.24 | | 150 | Accept | 0.21846 | 0.06502 | 0.083077 | 0.084345 | svm | BoxConstraint: 59.071 | | | | | | | | | KernelScale: 0.64404 | | 151 | Accept | 0.21231 | 0.058461 | 0.083077 | 0.084345 | svm | BoxConstraint: 27.03 | | | | | | | | | KernelScale: 0.58467 | | 152 | Accept | 0.16 | 0.054338 | 0.083077 | 0.084345 | svm | BoxConstraint: 1.0872 | | | | | | | | | KernelScale: 1.1762 | | 153 | Accept | 0.17846 | 0.055672 | 0.083077 | 0.084345 | svm | BoxConstraint: 0.55409 | | | | | | | | | KernelScale: 1.0383 | | 154 | Accept | 0.37231 | 0.056181 | 0.083077 | 0.084345 | svm | BoxConstraint: 0.002375 | | | | | | | | | KernelScale: 1.1605 | | 155 | Accept | 0.16923 | 0.057591 | 0.083077 | 0.084345 | svm | BoxConstraint: 0.5264 | | | | | | | | | KernelScale: 1.7249 | | 156 | Accept | 0.16308 | 0.076268 | 0.083077 | 0.084345 | svm | BoxConstraint: 0.4426 | | | | | | | | | KernelScale: 1.3849 | | 157 | Accept | 0.16308 | 0.071124 | 0.083077 | 0.084345 | svm | BoxConstraint: 0.43171 | | | | | | | | | KernelScale: 1.2674 | | 158 | Accept | 0.17231 | 0.059062 | 0.083077 | 0.084345 | svm | BoxConstraint: 0.33464 | | | | | | | | | KernelScale: 1.3315 | | 159 | Accept | 0.19385 | 0.069434 | 0.083077 | 0.084345 | svm | BoxConstraint: 0.15356 | | | | | | | | | KernelScale: 1.9858 | | 160 | Accept | 0.20308 | 0.057155 | 0.083077 | 0.084345 | svm | BoxConstraint: 0.1477 | | | | | | | | | KernelScale: 1.5045 | |=============================================================================================================================================| | Iter | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | result | loss | & validation (sec)| validation loss | validation loss | | | |=============================================================================================================================================| | 161 | Accept | 0.23692 | 0.062405 | 0.083077 | 0.084345 | svm | BoxConstraint: 0.42828 | | | | | | | | | KernelScale: 14.853 | | 162 | Accept | 0.086154 | 5.1922 | 0.083077 | 0.084288 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 275 | | | | | | | | | MinLeafSize: 47 | | 163 | Accept | 0.083077 | 5.2334 | 0.083077 | 0.084062 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 287 | | | | | | | | | MinLeafSize: 49 | | 164 | Accept | 0.20923 | 0.066873 | 0.083077 | 0.084062 | svm | BoxConstraint: 1.6705 | | | | | | | | | KernelScale: 0.51936 | | 165 | Accept | 0.083077 | 4.8932 | 0.083077 | 0.083699 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 274 | | | | | | | | | MinLeafSize: 48 | | 166 | Accept | 0.086154 | 5.2371 | 0.083077 | 0.083963 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 296 | | | | | | | | | MinLeafSize: 49 | | 167 | Accept | 0.37231 | 0.067167 | 0.083077 | 0.083963 | svm | BoxConstraint: 0.0024433 | | | | | | | | | KernelScale: 1.378 | | 168 | Accept | 0.32923 | 0.067094 | 0.083077 | 0.083963 | svm | BoxConstraint: 0.62588 | | | | | | | | | KernelScale: 0.49029 | | 169 | Accept | 0.19077 | 0.068847 | 0.083077 | 0.083963 | svm | BoxConstraint: 0.2494 | | | | | | | | | KernelScale: 1.2705 | | 170 | Accept | 0.24 | 0.055606 | 0.083077 | 0.083963 | svm | BoxConstraint: 0.85565 | | | | | | | | | KernelScale: 0.50742 | | 171 | Accept | 0.12923 | 3.9215 | 0.083077 | 0.083927 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 218 | | | | | | | | | MinLeafSize: 53 | | 172 | Accept | 0.11077 | 5.1249 | 0.083077 | 0.083981 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 283 | | | | | | | | | MinLeafSize: 34 | | 173 | Accept | 0.37231 | 6.449 | 0.083077 | 0.083855 | ensemble | Method: Bag | | | | | | | | | NumLearningCycles: 298 | | | | | | | | | MinLeafSize: 137 | | 174 | Accept | 0.20615 | 4.7277 | 0.083077 | 0.083837 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 263 | | | | | | | | | MinLeafSize: 128 | | 175 | Best | 0.083077 | 4.9307 | 0.083077 | 0.083627 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 274 | | | | | | | | | MinLeafSize: 44 | | 176 | Accept | 0.089231 | 4.9299 | 0.083077 | 0.083892 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 274 | | | | | | | | | MinLeafSize: 45 | | 177 | Accept | 0.092308 | 4.9475 | 0.083077 | 0.083668 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 274 | | | | | | | | | MinLeafSize: 52 | | 178 | Accept | 0.086154 | 5.237 | 0.083077 | 0.08391 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 291 | | | | | | | | | MinLeafSize: 49 | | 179 | Accept | 0.083077 | 4.8488 | 0.083077 | 0.08378 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 272 | | | | | | | | | MinLeafSize: 49 | | 180 | Accept | 0.083077 | 4.9095 | 0.083077 | 0.083649 | ensemble | Method: LogitBoost | | | | | | | | | NumLearningCycles: 274 | | | | | | | | | MinLeafSize: 48 | __________________________________________________________ Optimization completed. Total iterations: 180 Total elapsed time: 624.9642 seconds Total time for training and validation: 456.1613 seconds Best observed learner is an ensemble model with: Learner: ensemble Method: LogitBoost NumLearningCycles: 274 MinLeafSize: 44 Observed validation loss: 0.083077 Time for training and validation: 4.9307 seconds Best estimated learner (returned model) is an ensemble model with: Learner: ensemble Method: LogitBoost NumLearningCycles: 275 MinLeafSize: 49 Estimated validation loss: 0.083649 Estimated time for training and validation: 5.0351 seconds Documentation for fitcauto display
fitcauto によって返される最終的なモデルが、最適な推定学習器となります。モデルを返す前に、関数は学習データ全体 (carsTrain)、リストされている Learner (またはモデル) のタイプ、および表示されたハイパーパラメーター値を使用して、モデルの再学習を行います。
テスト セットのパフォーマンスの評価
テスト セットに対するモデルのパフォーマンスを評価します。
testAccuracy = 1 - loss(Mdl,carsTest,"Origin")testAccuracy = 0.9389
confusionchart(carsTest.Origin,predict(Mdl,carsTest))
fitcauto を使用し、個別の変数で指定した予測子データと応答データに基づいて、最適化されたハイパーパラメーターをもつ分類モデルを自動的に選択します。
データの読み込み
humanactivity データ セットを読み込みます。このデータ セットには、人間の次の 5 種類の身体動作についての 24,075 個の観測値が含まれています。座る (1)、立つ (2)、歩く (3)、走る (4)、踊る (5)。各観測値には、スマートフォンの加速度センサーによって測定された加速度データから抽出した 60 個の特徴量が含まれています。変数 feat には、24,075 個の観測値に関する 60 個の特徴量から成る予測子データ行列が格納され、応答変数 actid には、観測値の身体動作 ID が整数として格納されています。
load humanactivityデータの分割
データを学習セットとテスト セットに分割します。モデルの選択に観測値の 90% を使用し、fitcauto によって返された最終モデルの検証に観測値の 10% を使用します。cvpartition を使用して、観測値の 10% をテスト用に確保します。
rng("default") % For reproducibility of the partition c = cvpartition(actid,"Holdout",0.10); trainingIndices = training(c); % Indices for the training set XTrain = feat(trainingIndices,:); YTrain = actid(trainingIndices); testIndices = test(c); % Indices for the test set XTest = feat(testIndices,:); YTest = actid(testIndices);
fitcauto の実行
学習データを fitcauto に渡します。学習データ XTrain に含まれる観測値が 10,000 を超えるため、ベイズ最適化ではなく ASHA 最適化を使用します。関数 fitcauto は、さまざまなハイパーパラメーターの値をもつ適切なモデル (または学習器) のタイプを無作為に選択し、学習データの小さいサブセットでこのモデルに学習させ、適切に機能するモデルをプロモートして、プロモートされたモデルを徐々に大きくした学習データのセットで再学習させます。この関数は、最適な交差検証性能をもち、すべての学習データで再学習済みのモデル、および最適化の詳細を含む table を返します。最適化を並列実行するよう指定します (Parallel Computing Toolbox™ が必要)。
既定の設定では、fitcauto は、最適化のプロット、および最適化の結果の反復表示を提供します。これらの結果を解釈する方法の詳細については、Verbose の表示を参照してください。
options = struct("Optimizer","asha","UseParallel",true); [Mdl,OptimizationResults] = fitcauto(XTrain,YTrain,"HyperparameterOptimizationOptions",options);
Starting parallel pool (parpool) using the 'Processes' profile ... 10-May-2024 09:40:17: Job Queued. Waiting for parallel pool job with ID 1 to start ... Connected to parallel pool with 6 workers. Copying objective function to workers... Done copying objective function to workers. Learner types to explore: ensemble, knn, nb, net, svm, tree Total iterations (MaxObjectiveEvaluations): 595 Total time (MaxTime): Inf |====================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Training set | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | size | | | |====================================================================================================================================================| | 1 | 6 | Best | 0.056535 | 4.0512 | 0.056535 | 271 | tree | MinLeafSize: 1 | | 2 | 6 | Accept | 0.74165 | 4.8665 | 0.056535 | 271 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 1.545 | | | | | | | | | | LayerSizes: 156 | | 3 | 6 | Accept | 0.094794 | 3.0045 | 0.056535 | 271 | knn | NumNeighbors: 27 | | 4 | 6 | Accept | 0.74165 | 8.1847 | 0.056535 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 0.0019682 | | | | | | | | | | KernelScale: 87.187 | | 5 | 4 | Accept | 0.74165 | 8.8076 | 0.056535 | 271 | knn | NumNeighbors: 1634 | | 6 | 4 | Accept | 0.74165 | 8.2922 | 0.056535 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 0.012947 | | | | | | | | | | KernelScale: 45.176 | | 7 | 4 | Accept | 0.74165 | 8.4398 | 0.056535 | 271 | knn | NumNeighbors: 375 | | 8 | 6 | Accept | 0.057873 | 4.062 | 0.056535 | 271 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.0016149 | | | | | | | | | | LayerSizes: 18 | | 9 | 6 | Best | 0.049243 | 2.4566 | 0.049243 | 1084 | tree | MinLeafSize: 1 | | 10 | 6 | Accept | 0.062719 | 3.6253 | 0.049243 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 11 | 6 | Accept | 0.74165 | 8.1485 | 0.049243 | 271 | knn | NumNeighbors: 331 | | 12 | 6 | Accept | 0.74165 | 5.4441 | 0.049243 | 271 | knn | NumNeighbors: 3719 | | 13 | 6 | Best | 0.032075 | 13.099 | 0.032075 | 1084 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.0016149 | | | | | | | | | | LayerSizes: 18 | | 14 | 6 | Accept | 0.064934 | 1.1667 | 0.032075 | 271 | knn | NumNeighbors: 2 | | 15 | 6 | Accept | 0.050858 | 1.3991 | 0.032075 | 1084 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 16 | 6 | Accept | 0.74165 | 4.9367 | 0.032075 | 271 | knn | NumNeighbors: 2131 | | 17 | 6 | Accept | 0.038998 | 1.3761 | 0.032075 | 271 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 3.2054e-06 | | | | | | | | | | LayerSizes: 74 | | 18 | 6 | Accept | 0.12327 | 1.8297 | 0.032075 | 271 | knn | NumNeighbors: 34 | | 19 | 6 | Accept | 0.73929 | 12.116 | 0.032075 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 91.704 | | | | | | | | | | KernelScale: 0.015616 | | 20 | 6 | Accept | 0.17634 | 3.8994 | 0.032075 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.003312 | | | | | | | | | | KernelScale: 851.79 | |====================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Training set | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | size | | | |====================================================================================================================================================| | 21 | 6 | Accept | 0.74165 | 4.6497 | 0.032075 | 271 | knn | NumNeighbors: 678 | | 22 | 6 | Accept | 0.13587 | 2.3404 | 0.032075 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.019161 | | | | | | | | | | KernelScale: 153.3 | | 23 | 6 | Best | 0.026352 | 5.17 | 0.026352 | 1084 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 3.2054e-06 | | | | | | | | | | LayerSizes: 74 | | 24 | 6 | Accept | 0.74165 | 10.604 | 0.026352 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 0.0070102 | | | | | | | | | | KernelScale: 0.0038694 | | 25 | 6 | Accept | 0.043243 | 2.9641 | 0.026352 | 1084 | knn | NumNeighbors: 2 | | 26 | 6 | Best | 0.011861 | 29.125 | 0.011861 | 4334 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 3.2054e-06 | | | | | | | | | | LayerSizes: 74 | | 27 | 6 | Accept | 0.63379 | 66.627 | 0.011861 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 0.0028215 | | | | | | | | | | Standardize: true | | 28 | 6 | Accept | 0.74165 | 4.5204 | 0.011861 | 271 | knn | NumNeighbors: 1428 | | 29 | 6 | Accept | 0.05755 | 1.7987 | 0.011861 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 1.0426 | | | | | | | | | | KernelScale: 15.383 | | 30 | 6 | Accept | 0.054458 | 0.67873 | 0.011861 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 31 | 6 | Accept | 0.053904 | 0.66551 | 0.011861 | 1084 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 32 | 6 | Accept | 0.55727 | 91.003 | 0.011861 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 0.0059208 | | | | | | | | | | Standardize: false | | 33 | 6 | Accept | 0.74165 | 11.662 | 0.011861 | 271 | ensemble | Method: AdaBoostM2 | | | | | | | | | | NumLearningCycles: 270 | | | | | | | | | | MinLeafSize: 1014 | | | | | | | | | | MaxNumSplits: 27 | | 34 | 6 | Accept | 0.74165 | 10.501 | 0.011861 | 271 | ensemble | Method: AdaBoostM2 | | | | | | | | | | NumLearningCycles: 300 | | | | | | | | | | MinLeafSize: 184 | | | | | | | | | | MaxNumSplits: 18 | | 35 | 6 | Accept | 0.73842 | 168.5 | 0.011861 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 17.493 | | | | | | | | | | Standardize: true | | 36 | 6 | Accept | 0.73417 | 171.15 | 0.011861 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 856.62 | | | | | | | | | | Standardize: false | | 37 | 6 | Accept | 0.19139 | 173.58 | 0.011861 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 4.6757 | | | | | | | | | | Standardize: true | | 38 | 6 | Accept | 0.032536 | 3.7049 | 0.011861 | 1084 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 1.0426 | | | | | | | | | | KernelScale: 15.383 | | 39 | 6 | Accept | 0.74165 | 1.524 | 0.011861 | 271 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.29242 | | | | | | | | | | LayerSizes: [ 20 14 ] | | 40 | 6 | Accept | 0.052381 | 1.2116 | 0.011861 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | |====================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Training set | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | size | | | |====================================================================================================================================================| | 41 | 6 | Accept | 0.054966 | 0.65989 | 0.011861 | 1084 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 42 | 6 | Accept | 0.029952 | 18.516 | 0.011861 | 271 | ensemble | Method: AdaBoostM2 | | | | | | | | | | NumLearningCycles: 229 | | | | | | | | | | MinLeafSize: 7 | | | | | | | | | | MaxNumSplits: 31 | | 43 | 6 | Accept | 0.62415 | 4.6925 | 0.011861 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.019139 | | | | | | | | | | KernelScale: 0.86008 | | 44 | 6 | Accept | 0.069503 | 32.867 | 0.011861 | 271 | net | Activations: tanh | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.096e-06 | | | | | | | | | | LayerSizes: [ 289 2 ] | | 45 | 6 | Accept | 0.051366 | 1.214 | 0.011861 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 46 | 6 | Accept | 0.73888 | 11.397 | 0.011861 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 939.27 | | | | | | | | | | KernelScale: 0.0013942 | | 47 | 6 | Accept | 0.027691 | 34.472 | 0.011861 | 4334 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.0016149 | | | | | | | | | | LayerSizes: 18 | | 48 | 6 | Accept | 0.15341 | 0.51534 | 0.011861 | 271 | tree | MinLeafSize: 58 | | 49 | 6 | Accept | 0.32204 | 4.652 | 0.011861 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 4.6764 | | | | | | | | | | KernelScale: 2.4604 | | 50 | 6 | Accept | 0.11459 | 2.3006 | 0.011861 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 94.31 | | | | | | | | | | KernelScale: 944.78 | | 51 | 6 | Accept | 0.047351 | 0.96901 | 0.011861 | 1084 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 52 | 6 | Accept | 0.04412 | 0.66376 | 0.011861 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 53 | 6 | Accept | 0.048689 | 1.4012 | 0.011861 | 271 | knn | NumNeighbors: 1 | | 54 | 6 | Accept | 0.05492 | 0.73525 | 0.011861 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 55 | 6 | Accept | 0.73569 | 9.3198 | 0.011861 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.0031837 | | | | | | | | | | KernelScale: 0.29449 | | 56 | 6 | Accept | 0.053397 | 0.78515 | 0.011861 | 1084 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 57 | 6 | Accept | 0.11847 | 167.45 | 0.011861 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 2.7531 | | | | | | | | | | Standardize: true | | 58 | 6 | Accept | 0.049382 | 0.82672 | 0.011861 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 59 | 6 | Accept | 0.064242 | 1.1364 | 0.011861 | 271 | knn | NumNeighbors: 2 | | 60 | 6 | Accept | 0.076565 | 0.88796 | 0.011861 | 271 | tree | MinLeafSize: 2 | |====================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Training set | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | size | | | |====================================================================================================================================================| | 61 | 6 | Accept | 0.059904 | 0.94055 | 0.011861 | 271 | knn | NumNeighbors: 1 | | 62 | 6 | Accept | 0.033875 | 3.2696 | 0.011861 | 1084 | knn | NumNeighbors: 1 | | 63 | 6 | Accept | 0.066181 | 9.5233 | 0.011861 | 271 | net | Activations: relu | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 3.0708e-06 | | | | | | | | | | LayerSizes: [ 85 6 13 ] | | 64 | 6 | Accept | 0.71045 | 6.4658 | 0.011861 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 86.648 | | | | | | | | | | KernelScale: 0.68451 | | 65 | 6 | Accept | 0.015368 | 27.396 | 0.011861 | 1084 | ensemble | Method: AdaBoostM2 | | | | | | | | | | NumLearningCycles: 229 | | | | | | | | | | MinLeafSize: 7 | | | | | | | | | | MaxNumSplits: 31 | | 66 | 6 | Accept | 0.73906 | 11.212 | 0.011861 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 1.622 | | | | | | | | | | KernelScale: 0.0038764 | | 67 | 6 | Accept | 0.053904 | 0.66704 | 0.011861 | 1084 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 68 | 6 | Accept | 0.73911 | 11.194 | 0.011861 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 49.928 | | | | | | | | | | KernelScale: 0.0035535 | | 69 | 6 | Accept | 0.74165 | 0.19475 | 0.011861 | 271 | tree | MinLeafSize: 1365 | | 70 | 6 | Accept | 0.046751 | 3.2493 | 0.011861 | 271 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 1.0454e-06 | | | | | | | | | | LayerSizes: [ 4 7 26 ] | | 71 | 6 | Best | 0.0077995 | 37.224 | 0.0077995 | 4334 | ensemble | Method: AdaBoostM2 | | | | | | | | | | NumLearningCycles: 229 | | | | | | | | | | MinLeafSize: 7 | | | | | | | | | | MaxNumSplits: 31 | | 72 | 6 | Accept | 0.096963 | 3.5085 | 0.0077995 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 8.0828 | | | | | | | | | | KernelScale: 171.49 | | 73 | 6 | Accept | 0.74165 | 168.14 | 0.0077995 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 596 | | | | | | | | | | Standardize: true | | 74 | 6 | Accept | 0.057227 | 0.76102 | 0.0077995 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 75 | 6 | Accept | 0.059304 | 14.374 | 0.0077995 | 1084 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 1.0454e-06 | | | | | | | | | | LayerSizes: [ 4 7 26 ] | | 76 | 6 | Accept | 0.73975 | 11.338 | 0.0077995 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.56028 | | | | | | | | | | KernelScale: 0.0093903 | | 77 | 6 | Accept | 0.057966 | 0.71414 | 0.0077995 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 78 | 6 | Accept | 0.046151 | 0.7308 | 0.0077995 | 1084 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 79 | 6 | Accept | 0.048505 | 173.22 | 0.0077995 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 0.61252 | | | | | | | | | | Standardize: true | | 80 | 6 | Accept | 0.068257 | 20.884 | 0.0077995 | 271 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.0891e-05 | | | | | | | | | | LayerSizes: [ 80 86 2 ] | |====================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Training set | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | size | | | |====================================================================================================================================================| | 81 | 6 | Accept | 0.74165 | 4.8285 | 0.0077995 | 271 | knn | NumNeighbors: 1560 | | 82 | 6 | Accept | 0.72946 | 168.86 | 0.0077995 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 754.46 | | | | | | | | | | Standardize: false | | 83 | 6 | Accept | 0.3807 | 169.46 | 0.0077995 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 44.255 | | | | | | | | | | Standardize: false | | 84 | 6 | Accept | 0.081549 | 130.35 | 0.0077995 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 0.098769 | | | | | | | | | | Standardize: true | | 85 | 6 | Accept | 0.74165 | 3.7268 | 0.0077995 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 8.8662 | | | | | | | | | | KernelScale: 541.83 | | 86 | 6 | Accept | 0.10033 | 0.66666 | 0.0077995 | 271 | tree | MinLeafSize: 17 | | 87 | 6 | Accept | 0.25448 | 171.99 | 0.0077995 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 7.1617 | | | | | | | | | | Standardize: true | | 88 | 6 | Accept | 0.05755 | 0.71855 | 0.0077995 | 1084 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 89 | 6 | Accept | 0.049105 | 1.283 | 0.0077995 | 271 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 4.2009e-08 | | | | | | | | | | LayerSizes: 133 | | 90 | 6 | Accept | 0.072457 | 0.73998 | 0.0077995 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 91 | 6 | Accept | 0.021922 | 6.709 | 0.0077995 | 4334 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 1.0426 | | | | | | | | | | KernelScale: 15.383 | | 92 | 6 | Accept | 0.73883 | 12.629 | 0.0077995 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.19021 | | | | | | | | | | KernelScale: 0.13128 | | 93 | 6 | Accept | 0.74165 | 10.879 | 0.0077995 | 271 | ensemble | Method: AdaBoostM2 | | | | | | | | | | NumLearningCycles: 288 | | | | | | | | | | MinLeafSize: 4760 | | | | | | | | | | MaxNumSplits: 25 | | 94 | 6 | Accept | 0.030967 | 3.5848 | 0.0077995 | 1084 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 4.2009e-08 | | | | | | | | | | LayerSizes: 133 | | 95 | 6 | Accept | 0.067842 | 0.71485 | 0.0077995 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 96 | 6 | Accept | 0.74165 | 0.21973 | 0.0077995 | 271 | tree | MinLeafSize: 5731 | | 97 | 6 | Accept | 0.51435 | 97.511 | 0.0077995 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 0.010292 | | | | | | | | | | Standardize: false | | 98 | 6 | Accept | 0.049474 | 0.7033 | 0.0077995 | 1084 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 99 | 6 | Accept | 0.078503 | 3.5346 | 0.0077995 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 65.734 | | | | | | | | | | KernelScale: 4.5947 | | 100 | 6 | Accept | 0.047628 | 0.73303 | 0.0077995 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | |====================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Training set | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | size | | | |====================================================================================================================================================| | 101 | 6 | Accept | 0.037428 | 7.6392 | 0.0077995 | 271 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 6.2668e-05 | | | | | | | | | | LayerSizes: [ 171 9 ] | | 102 | 6 | Accept | 0.73962 | 9.4795 | 0.0077995 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 28.585 | | | | | | | | | | KernelScale: 0.14053 | | 103 | 6 | Accept | 0.74165 | 164.68 | 0.0077995 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 209.59 | | | | | | | | | | Standardize: true | | 104 | 6 | Accept | 0.026814 | 28.779 | 0.0077995 | 1084 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 6.2668e-05 | | | | | | | | | | LayerSizes: [ 171 9 ] | | 105 | 6 | Best | 0.0053074 | 62.586 | 0.0053074 | 17335 | ensemble | Method: AdaBoostM2 | | | | | | | | | | NumLearningCycles: 229 | | | | | | | | | | MinLeafSize: 7 | | | | | | | | | | MaxNumSplits: 31 | | 106 | 6 | Accept | 0.74792 | 1.5953 | 0.0053074 | 271 | ensemble | Method: AdaBoostM2 | | | | | | | | | | NumLearningCycles: 238 | | | | | | | | | | MinLeafSize: 1 | | | | | | | | | | MaxNumSplits: 57 | | 107 | 6 | Accept | 0.74165 | 0.3562 | 0.0053074 | 271 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.6829 | | | | | | | | | | LayerSizes: [ 40 4 ] | | 108 | 6 | Accept | 0.060227 | 3.6079 | 0.0053074 | 271 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.0069133 | | | | | | | | | | LayerSizes: [ 3 12 12 ] | | 109 | 6 | Accept | 0.049658 | 0.70789 | 0.0053074 | 1084 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 110 | 6 | Accept | 0.59156 | 79.711 | 0.0053074 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 0.0033878 | | | | | | | | | | Standardize: false | | 111 | 6 | Accept | 0.74165 | 5.2086 | 0.0053074 | 271 | knn | NumNeighbors: 373 | | 112 | 6 | Accept | 0.047766 | 0.83263 | 0.0053074 | 271 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 2.319e-05 | | | | | | | | | | LayerSizes: 17 | | 113 | 6 | Accept | 0.054089 | 0.72423 | 0.0053074 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 114 | 6 | Accept | 0.02866 | 3.1959 | 0.0053074 | 1084 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 2.319e-05 | | | | | | | | | | LayerSizes: 17 | | 115 | 6 | Accept | 0.74165 | 4.4465 | 0.0053074 | 271 | knn | NumNeighbors: 1189 | | 116 | 6 | Accept | 0.74165 | 4.442 | 0.0053074 | 271 | knn | NumNeighbors: 1145 | | 117 | 6 | Accept | 0.032444 | 18.279 | 0.0053074 | 271 | ensemble | Method: Bag | | | | | | | | | | NumLearningCycles: 268 | | | | | | | | | | MinLeafSize: 1 | | | | | | | | | | MaxNumSplits: 79 | | 118 | 6 | Accept | 0.069919 | 0.70972 | 0.0053074 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 119 | 6 | Accept | 0.098025 | 155.9 | 0.0053074 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 6.0379 | | | | | | | | | | Standardize: false | | 120 | 6 | Accept | 0.14745 | 161.41 | 0.0053074 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 2.9595 | | | | | | | | | | Standardize: true | |====================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Training set | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | size | | | |====================================================================================================================================================| | 121 | 6 | Accept | 0.046243 | 0.6139 | 0.0053074 | 271 | net | Activations: relu | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 6.6757e-07 | | | | | | | | | | LayerSizes: 65 | | 122 | 6 | Accept | 0.053766 | 2.3698 | 0.0053074 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 914.18 | | | | | | | | | | KernelScale: 85.472 | | 123 | 6 | Accept | 0.025752 | 1.2539 | 0.0053074 | 1084 | net | Activations: relu | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 6.6757e-07 | | | | | | | | | | LayerSizes: 65 | | 124 | 6 | Accept | 0.73925 | 10.728 | 0.0053074 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 544.31 | | | | | | | | | | KernelScale: 0.007687 | | 125 | 6 | Accept | 0.054966 | 14.621 | 0.0053074 | 271 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.9041e-06 | | | | | | | | | | LayerSizes: [ 210 9 ] | | 126 | 6 | Accept | 0.02806 | 25.852 | 0.0053074 | 1084 | ensemble | Method: Bag | | | | | | | | | | NumLearningCycles: 268 | | | | | | | | | | MinLeafSize: 1 | | | | | | | | | | MaxNumSplits: 79 | | 127 | 6 | Accept | 0.04412 | 1.0243 | 0.0053074 | 271 | tree | MinLeafSize: 2 | | 128 | 6 | Accept | 0.74165 | 4.9624 | 0.0053074 | 271 | knn | NumNeighbors: 2336 | | 129 | 6 | Accept | 0.12244 | 2.6488 | 0.0053074 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.045281 | | | | | | | | | | KernelScale: 593.59 | | 130 | 6 | Accept | 0.030321 | 15.349 | 0.0053074 | 271 | ensemble | Method: Bag | | | | | | | | | | NumLearningCycles: 205 | | | | | | | | | | MinLeafSize: 1 | | | | | | | | | | MaxNumSplits: 76 | | 131 | 5 | Accept | 0.077441 | 1.2571 | 0.0053074 | 271 | knn | NumNeighbors: 6 | | 132 | 5 | Accept | 0.065811 | 0.75844 | 0.0053074 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 133 | 6 | Accept | 0.05492 | 2.1752 | 0.0053074 | 1084 | tree | MinLeafSize: 2 | | 134 | 6 | Accept | 0.068303 | 1.1461 | 0.0053074 | 271 | knn | NumNeighbors: 6 | | 135 | 6 | Accept | 0.74165 | 9.9296 | 0.0053074 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 0.012836 | | | | | | | | | | KernelScale: 0.0013573 | | 136 | 6 | Accept | 0.048182 | 3.954 | 0.0053074 | 271 | net | Activations: tanh | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.0917e-09 | | | | | | | | | | LayerSizes: 27 | | 137 | 6 | Accept | 0.74165 | 4.8146 | 0.0053074 | 271 | knn | NumNeighbors: 5959 | | 138 | 6 | Accept | 0.027368 | 20.876 | 0.0053074 | 1084 | ensemble | Method: Bag | | | | | | | | | | NumLearningCycles: 205 | | | | | | | | | | MinLeafSize: 1 | | | | | | | | | | MaxNumSplits: 76 | | 139 | 6 | Accept | 0.037936 | 1.6741 | 0.0053074 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 3.9775 | | | | | | | | | | KernelScale: 13.747 | | 140 | 6 | Accept | 0.10181 | 2.2241 | 0.0053074 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.5329 | | | | | | | | | | KernelScale: 458.88 | |====================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Training set | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | size | | | |====================================================================================================================================================| | 141 | 6 | Accept | 0.066688 | 164.37 | 0.0053074 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 1.1777 | | | | | | | | | | Standardize: true | | 142 | 6 | Accept | 0.049474 | 14.68 | 0.0053074 | 1084 | net | Activations: tanh | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.0917e-09 | | | | | | | | | | LayerSizes: 27 | | 143 | 6 | Accept | 0.076611 | 8.3053 | 0.0053074 | 271 | net | Activations: relu | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.2759e-05 | | | | | | | | | | LayerSizes: [ 71 14 ] | | 144 | 6 | Accept | 0.012322 | 39.233 | 0.0053074 | 4334 | net | Activations: relu | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 6.6757e-07 | | | | | | | | | | LayerSizes: 65 | | 145 | 6 | Accept | 0.74165 | 4.99 | 0.0053074 | 271 | knn | NumNeighbors: 980 | | 146 | 6 | Accept | 0.026445 | 2.363 | 0.0053074 | 1084 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 3.9775 | | | | | | | | | | KernelScale: 13.747 | | 147 | 6 | Accept | 0.04209 | 2.3578 | 0.0053074 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 10.427 | | | | | | | | | | KernelScale: 16.103 | | 148 | 6 | Accept | 0.069688 | 0.70265 | 0.0053074 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 149 | 6 | Accept | 0.74165 | 4.537 | 0.0053074 | 271 | knn | NumNeighbors: 767 | | 150 | 6 | Accept | 0.023214 | 2.3517 | 0.0053074 | 1084 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 10.427 | | | | | | | | | | KernelScale: 16.103 | | 151 | 6 | Accept | 0.73929 | 9.8139 | 0.0053074 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 6.3626 | | | | | | | | | | KernelScale: 0.002014 | | 152 | 6 | Accept | 0.011215 | 176.8 | 0.0053074 | 4334 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 6.2668e-05 | | | | | | | | | | LayerSizes: [ 171 9 ] | | 153 | 6 | Accept | 0.011769 | 3.7526 | 0.0053074 | 4334 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 10.427 | | | | | | | | | | KernelScale: 16.103 | | 154 | 6 | Accept | 0.74165 | 0.17534 | 0.0053074 | 271 | tree | MinLeafSize: 3138 | | 155 | 6 | Accept | 0.74165 | 10.965 | 0.0053074 | 271 | ensemble | Method: Bag | | | | | | | | | | NumLearningCycles: 265 | | | | | | | | | | MinLeafSize: 272 | | | | | | | | | | MaxNumSplits: 32 | | 156 | 6 | Accept | 0.74165 | 9.6757 | 0.0053074 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 0.0027945 | | | | | | | | | | KernelScale: 0.047512 | | 157 | 6 | Accept | 0.055058 | 11.792 | 0.0053074 | 271 | net | Activations: relu | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.0004489 | | | | | | | | | | LayerSizes: [ 127 10 ] | | 158 | 6 | Accept | 0.028014 | 2.2151 | 0.0053074 | 1084 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 914.18 | | | | | | | | | | KernelScale: 85.472 | | 159 | 6 | Accept | 0.044167 | 0.74389 | 0.0053074 | 271 | net | Activations: relu | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 1.267e-06 | | | | | | | | | | LayerSizes: [ 21 51 ] | | 160 | 6 | Accept | 0.74165 | 4.4576 | 0.0053074 | 271 | knn | NumNeighbors: 1090 | |====================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Training set | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | size | | | |====================================================================================================================================================| | 161 | 6 | Accept | 0.40027 | 4.8502 | 0.0053074 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 24.409 | | | | | | | | | | KernelScale: 1.6395 | | 162 | 6 | Accept | 0.022199 | 1.9746 | 0.0053074 | 1084 | net | Activations: relu | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 1.267e-06 | | | | | | | | | | LayerSizes: [ 21 51 ] | | 163 | 6 | Accept | 0.043013 | 292.66 | 0.0053074 | 1084 | nb | DistributionNames: kernel | | | | | | | | | | Width: 0.61252 | | | | | | | | | | Standardize: true | | 164 | 6 | Accept | 0.74165 | 8.3033 | 0.0053074 | 271 | ensemble | Method: Bag | | | | | | | | | | NumLearningCycles: 216 | | | | | | | | | | MinLeafSize: 8895 | | | | | | | | | | MaxNumSplits: 54 | | 165 | 6 | Accept | 0.060412 | 0.74162 | 0.0053074 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 166 | 6 | Accept | 0.74165 | 4.6892 | 0.0053074 | 271 | knn | NumNeighbors: 4825 | | 167 | 6 | Accept | 0.053812 | 1.6486 | 0.0053074 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 41.59 | | | | | | | | | | KernelScale: 66.838 | | 168 | 6 | Accept | 0.73934 | 9.6965 | 0.0053074 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 3.5182 | | | | | | | | | | KernelScale: 0.0021656 | | 169 | 6 | Accept | 0.040844 | 0.65233 | 0.0053074 | 271 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 3.4639e-08 | | | | | | | | | | LayerSizes: 12 | | 170 | 6 | Accept | 0.65807 | 0.50924 | 0.0053074 | 271 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.2677 | | | | | | | | | | LayerSizes: 46 | | 171 | 6 | Accept | 0.028844 | 2.7895 | 0.0053074 | 1084 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 41.59 | | | | | | | | | | KernelScale: 66.838 | | 172 | 6 | Accept | 0.041213 | 2.536 | 0.0053074 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 114.59 | | | | | | | | | | KernelScale: 169.23 | | 173 | 6 | Accept | 0.064981 | 7.9256 | 0.0053074 | 271 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 6.46e-10 | | | | | | | | | | LayerSizes: [ 94 6 ] | | 174 | 6 | Accept | 0.033459 | 1.8565 | 0.0053074 | 1084 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 3.4639e-08 | | | | | | | | | | LayerSizes: 12 | | 175 | 6 | Accept | 0.059904 | 0.80685 | 0.0053074 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 176 | 6 | Accept | 0.28761 | 1.5076 | 0.0053074 | 271 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 9.7994e-09 | | | | | | | | | | LayerSizes: 1 | | 177 | 6 | Accept | 0.73957 | 11.525 | 0.0053074 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.01816 | | | | | | | | | | KernelScale: 0.0019383 | | 178 | 6 | Accept | 0.1319 | 2.3554 | 0.0053074 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.065299 | | | | | | | | | | KernelScale: 930.44 | | 179 | 6 | Accept | 0.027829 | 2.7788 | 0.0053074 | 1084 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 114.59 | | | | | | | | | | KernelScale: 169.23 | | 180 | 6 | Accept | 0.07338 | 0.72933 | 0.0053074 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | |====================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Training set | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | size | | | |====================================================================================================================================================| | 181 | 6 | Accept | 0.74165 | 4.4535 | 0.0053074 | 271 | knn | NumNeighbors: 232 | | 182 | 6 | Accept | 0.059119 | 0.77186 | 0.0053074 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 183 | 6 | Accept | 0.050535 | 0.69862 | 0.0053074 | 1084 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 184 | 6 | Accept | 0.74165 | 10.098 | 0.0053074 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 0.0028005 | | | | | | | | | | KernelScale: 0.010353 | | 185 | 6 | Accept | 0.046243 | 0.71168 | 0.0053074 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 186 | 6 | Accept | 0.74165 | 4.6281 | 0.0053074 | 271 | knn | NumNeighbors: 857 | | 187 | 6 | Accept | 0.1817 | 2.3545 | 0.0053074 | 271 | knn | NumNeighbors: 66 | | 188 | 6 | Accept | 0.051551 | 0.70548 | 0.0053074 | 1084 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 189 | 6 | Accept | 0.74165 | 4.6952 | 0.0053074 | 271 | knn | NumNeighbors: 6918 | | 190 | 6 | Accept | 0.04329 | 1.515 | 0.0053074 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 900.92 | | | | | | | | | | KernelScale: 44.444 | | 191 | 6 | Accept | 0.72199 | 6.2291 | 0.0053074 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 28.613 | | | | | | | | | | KernelScale: 0.4911 | | 192 | 6 | Accept | 0.043936 | 78.329 | 0.0053074 | 271 | ensemble | Coding: onevsone | | | | | | | | | | Method: LogitBoost | | | | | | | | | | NumLearningCycles: 211 | | | | | | | | | | MinLeafSize: 4 | | | | | | | | | | MaxNumSplits: 13 | | 193 | 6 | Accept | 0.036275 | 0.85738 | 0.0053074 | 271 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 5.0779e-06 | | | | | | | | | | LayerSizes: 31 | | 194 | 6 | Accept | 0.014353 | 6.0141 | 0.0053074 | 4334 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 3.9775 | | | | | | | | | | KernelScale: 13.747 | | 195 | 6 | Accept | 0.039921 | 0.67116 | 0.0053074 | 271 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.00035338 | | | | | | | | | | LayerSizes: 4 | | 196 | 6 | Accept | 0.027552 | 2.0495 | 0.0053074 | 1084 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 900.92 | | | | | | | | | | KernelScale: 44.444 | | 197 | 6 | Accept | 0.52917 | 0.28918 | 0.0053074 | 271 | tree | MinLeafSize: 116 | | 198 | 6 | Accept | 0.083718 | 1.4053 | 0.0053074 | 271 | knn | NumNeighbors: 12 | | 199 | 6 | Accept | 0.74165 | 0.63448 | 0.0053074 | 271 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.27358 | | | | | | | | | | LayerSizes: [ 21 58 47 ] | | 200 | 6 | Accept | 0.068903 | 2.0491 | 0.0053074 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.53387 | | | | | | | | | | KernelScale: 9.7797 | |====================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Training set | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | size | | | |====================================================================================================================================================| | 201 | 6 | Accept | 0.023675 | 2.9905 | 0.0053074 | 1084 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 5.0779e-06 | | | | | | | | | | LayerSizes: 31 | | 202 | 6 | Accept | 0.74165 | 81.003 | 0.0053074 | 271 | ensemble | Coding: onevsone | | | | | | | | | | Method: LogitBoost | | | | | | | | | | NumLearningCycles: 242 | | | | | | | | | | MinLeafSize: 2550 | | | | | | | | | | MaxNumSplits: 42 | | 203 | 6 | Accept | 0.74165 | 4.7264 | 0.0053074 | 271 | knn | NumNeighbors: 911 | | 204 | 6 | Accept | 0.031706 | 2.5716 | 0.0053074 | 1084 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.00035338 | | | | | | | | | | LayerSizes: 4 | | 205 | 6 | Accept | 0.058012 | 0.75697 | 0.0053074 | 271 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 206 | 6 | Accept | 0.013338 | 54.224 | 0.0053074 | 4334 | net | Activations: relu | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 1.267e-06 | | | | | | | | | | LayerSizes: [ 21 51 ] | | 207 | 6 | Accept | 0.74165 | 0.54395 | 0.0053074 | 271 | net | Activations: tanh | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.40394 | | | | | | | | | | LayerSizes: [ 23 93 ] | | 208 | 6 | Accept | 0.10167 | 2.3728 | 0.0053074 | 271 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 3.9533e-09 | | | | | | | | | | LayerSizes: [ 1 1 ] | | 209 | 6 | Accept | 0.74165 | 4.6078 | 0.0053074 | 271 | knn | NumNeighbors: 862 | | 210 | 6 | Accept | 0.36815 | 3.3522 | 0.0053074 | 271 | knn | NumNeighbors: 140 | | 211 | 6 | Accept | 0.74165 | 4.4738 | 0.0053074 | 271 | knn | NumNeighbors: 8006 | | 212 | 6 | Accept | 0.044997 | 43.489 | 0.0053074 | 271 | net | Activations: tanh | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 2.5417e-08 | | | | | | | | | | LayerSizes: [ 44 116 226 ] | | 213 | 6 | Accept | 0.05132 | 3.4248 | 0.0053074 | 271 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.0027776 | | | | | | | | | | LayerSizes: [ 72 8 ] | | 214 | 6 | Accept | 0.74165 | 45.286 | 0.0053074 | 271 | ensemble | Coding: onevsall | | | | | | | | | | Method: LogitBoost | | | | | | | | | | NumLearningCycles: 268 | | | | | | | | | | MinLeafSize: 2870 | | | | | | | | | | MaxNumSplits: 91 | | 215 | 6 | Accept | 0.7398 | 9.7726 | 0.0053074 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 68.727 | | | | | | | | | | KernelScale: 0.025309 | | 216 | 6 | Accept | 0.74165 | 4.64 | 0.0053074 | 271 | knn | NumNeighbors: 4617 | | 217 | 6 | Accept | 0.66781 | 4.5932 | 0.0053074 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.45784 | | | | | | | | | | KernelScale: 0.67635 | | 218 | 6 | Accept | 0.046843 | 5.8551 | 0.0053074 | 1084 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.0027776 | | | | | | | | | | LayerSizes: [ 72 8 ] | | 219 | 6 | Accept | 0.013153 | 26.521 | 0.0053074 | 4334 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 5.0779e-06 | | | | | | | | | | LayerSizes: 31 | | 220 | 6 | Accept | 0.74165 | 0.13889 | 0.0053074 | 271 | tree | MinLeafSize: 214 | |====================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Training set | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | size | | | |====================================================================================================================================================| | 221 | 6 | Accept | 0.042828 | 3.9304 | 0.0053074 | 271 | net | Activations: tanh | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.00032581 | | | | | | | | | | LayerSizes: 10 | | 222 | 6 | Accept | 0.74165 | 9.8725 | 0.0053074 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 0.32491 | | | | | | | | | | KernelScale: 0.0011646 | | 223 | 6 | Accept | 0.74165 | 0.12981 | 0.0053074 | 271 | tree | MinLeafSize: 405 | | 224 | 6 | Accept | 0.036829 | 9.849 | 0.0053074 | 1084 | net | Activations: tanh | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.00032581 | | | | | | | | | | LayerSizes: 10 | | 225 | 6 | Accept | 0.10001 | 1.5959 | 0.0053074 | 271 | knn | NumNeighbors: 24 | | 226 | 6 | Accept | 0.11805 | 0.42103 | 0.0053074 | 271 | tree | MinLeafSize: 51 | | 227 | 6 | Accept | 0.029444 | 93.524 | 0.0053074 | 1084 | ensemble | Coding: onevsone | | | | | | | | | | Method: LogitBoost | | | | | | | | | | NumLearningCycles: 211 | | | | | | | | | | MinLeafSize: 4 | | | | | | | | | | MaxNumSplits: 13 | | 228 | 6 | Accept | 0.74165 | 0.19998 | 0.0053074 | 271 | tree | MinLeafSize: 2088 | | 229 | 6 | Accept | 0.7416 | 168.62 | 0.0053074 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 20.609 | | | | | | | | | | Standardize: true | | 230 | 6 | Accept | 0.032906 | 151.18 | 0.0053074 | 1084 | net | Activations: tanh | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 2.5417e-08 | | | | | | | | | | LayerSizes: [ 44 116 226 ] | | 231 | 6 | Accept | 0.6787 | 70.071 | 0.0053074 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 0.0015605 | | | | | | | | | | Standardize: false | | 232 | 6 | Accept | 0.12955 | 3.7371 | 0.0053074 | 271 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 2.7261e-07 | | | | | | | | | | LayerSizes: 5 | | 233 | 6 | Accept | 0.52838 | 0.89457 | 0.0053074 | 271 | tree | MinLeafSize: 105 | | 234 | 6 | Accept | 0.055474 | 0.95035 | 0.0053074 | 271 | knn | NumNeighbors: 3 | | 235 | 6 | Accept | 0.74165 | 9.9179 | 0.0053074 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 0.2143 | | | | | | | | | | KernelScale: 0.19817 | | 236 | 6 | Accept | 0.032029 | 35.514 | 0.0053074 | 1084 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.9041e-06 | | | | | | | | | | LayerSizes: [ 210 9 ] | | 237 | 6 | Accept | 0.34696 | 8.8977 | 0.0053074 | 271 | net | Activations: relu | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.0022429 | | | | | | | | | | LayerSizes: [ 100 1 ] | | 238 | 6 | Accept | 0.28452 | 95.413 | 0.0053074 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 0.016256 | | | | | | | | | | Standardize: true | | 239 | 6 | Accept | 0.74165 | 171.77 | 0.0053074 | 271 | nb | DistributionNames: kernel | | | | | | | | | | Width: 132.12 | | | | | | | | | | Standardize: true | | 240 | 6 | Accept | 0.72877 | 7.0112 | 0.0053074 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 80.2 | | | | | | | | | | KernelScale: 0.4411 | |====================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Training set | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | size | | | |====================================================================================================================================================| | 241 | 6 | Accept | 0.03429 | 3.2421 | 0.0053074 | 1084 | knn | NumNeighbors: 3 | | 242 | 6 | Accept | 0.13144 | 8.5231 | 0.0053074 | 271 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 4.1361e-06 | | | | | | | | | | LayerSizes: [ 1 102 ] | | 243 | 6 | Accept | 0.058058 | 4.1342 | 0.0053074 | 271 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.00060115 | | | | | | | | | | LayerSizes: [ 3 8 ] | | 244 | 6 | Accept | 0.036598 | 33.658 | 0.0053074 | 1084 | net | Activations: relu | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.0004489 | | | | | | | | | | LayerSizes: [ 127 10 ] | | 245 | 6 | Accept | 0.12756 | 3.8626 | 0.0053074 | 271 | net | Activations: tanh | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.0034719 | | | | | | | | | | LayerSizes: [ 2 10 ] | | 246 | 6 | Accept | 0.05012 | 0.98248 | 0.0053074 | 1084 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 247 | 6 | Accept | 0.064842 | 0.86246 | 0.0053074 | 271 | tree | MinLeafSize: 6 | | 248 | 6 | Accept | 0.32126 | 16.422 | 0.0053074 | 271 | ensemble | Method: Bag | | | | | | | | | | NumLearningCycles: 287 | | | | | | | | | | MinLeafSize: 92 | | | | | | | | | | MaxNumSplits: 34 | | 249 | 6 | Accept | 0.74165 | 3.7533 | 0.0053074 | 271 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 0.0020576 | | | | | | | | | | KernelScale: 601.64 | | 250 | 6 | Accept | 0.73985 | 10.991 | 0.0053074 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.021505 | | | | | | | | | | KernelScale: 0.0044067 | | 251 | 6 | Accept | 0.01846 | 46.153 | 0.0053074 | 4334 | ensemble | Method: Bag | | | | | | | | | | NumLearningCycles: 205 | | | | | | | | | | MinLeafSize: 1 | | | | | | | | | | MaxNumSplits: 76 | | 252 | 6 | Accept | 0.12936 | 8.0856 | 0.0053074 | 1084 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.00060115 | | | | | | | | | | LayerSizes: [ 3 8 ] | | 253 | 6 | Accept | 0.013707 | 10.948 | 0.0053074 | 4334 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 900.92 | | | | | | | | | | KernelScale: 44.444 | | 254 | 6 | Accept | 0.30598 | 20.762 | 0.0053074 | 271 | net | Activations: tanh | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.5659e-09 | | | | | | | | | | LayerSizes: [ 173 2 ] | | 255 | 6 | Accept | 0.48768 | 4.4372 | 0.0053074 | 271 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 3.1297 | | | | | | | | | | KernelScale: 1.1348 | | 256 | 6 | Accept | 0.006415 | 11.83 | 0.0053074 | 17335 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 10.427 | | | | | | | | | | KernelScale: 16.103 | | 257 | 6 | Accept | 0.064427 | 0.76725 | 0.0053074 | 271 | nb | DistributionNames: ...
__________________________________________________________ Optimization completed. Total iterations: 595 Total elapsed time: 2507.5379 seconds Total time for training and validation: 14243.9529 seconds Best observed learner is a net model with: Learner: net Activations: sigmoid Standardize: true Lambda: 3.2054e-06 LayerSizes: 74 Observed validation loss: 0.0043843 Time for training and validation: 193.32 seconds Documentation for fitcauto display
fitcauto によって返される最終的なモデルが、観測された最適な学習器となります。モデルを返す前に、関数はすべての学習データ (XTrain と YTrain)、リストされている Learner (またはモデル) のタイプ、および表示されたハイパーパラメーター値を使用して、モデルの再学習を行います。
テスト セットのパフォーマンスの評価
テスト データ セットに対する最終モデルのパフォーマンスを評価します。
testAccuracy = 1 - loss(Mdl,XTest,YTest)
testAccuracy = 0.9992
最終モデルは、99% を超える観測値を正しく分類しています。
fitcauto を使用し、table で指定した予測子データと応答データに基づいて、最適化されたハイパーパラメーターをもつ分類モデルを自動的に選択します。データを fitcauto に渡す前に特徴選択を実行して、重要でない予測子をデータ セットから削除します。
データの読み込みと分割
標本ファイル CreditRating_Historical.dat を table に読み取ります。予測子データは、法人顧客リストの財務比率と業種の情報で構成されます。応答変数は、格付機関が割り当てた格付けから構成されます。データ セットの最初の数行をプレビューします。
creditrating = readtable("CreditRating_Historical.dat");
head(creditrating) ID WC_TA RE_TA EBIT_TA MVE_BVTD S_TA Industry Rating
_____ ______ ______ _______ ________ _____ ________ _______
62394 0.013 0.104 0.036 0.447 0.142 3 {'BB' }
48608 0.232 0.335 0.062 1.969 0.281 8 {'A' }
42444 0.311 0.367 0.074 1.935 0.366 1 {'A' }
48631 0.194 0.263 0.062 1.017 0.228 4 {'BBB'}
43768 0.121 0.413 0.057 3.647 0.466 12 {'AAA'}
39255 -0.117 -0.799 0.01 0.179 0.082 4 {'CCC'}
62236 0.087 0.158 0.049 0.816 0.324 2 {'BBB'}
39354 0.005 0.181 0.034 2.597 0.388 7 {'AA' }
変数 ID の各値は一意の顧客 ID であるため (つまり、length(unique(creditrating.ID)) は creditrating に含まれる観測値の数に等しい)、変数 ID は予測子としては適切ではありません。変数 ID を table から削除し、変数 Industry を categorical 変数に変換します。
creditrating = removevars(creditrating,"ID");
creditrating.Industry = categorical(creditrating.Industry);データを学習セットとテスト セットに分割します。モデル選択とハイパーパラメーター調整のプロセスに観測値の約 85% を使用し、fitcauto によって返された、新しいデータに対する最終モデルのパフォーマンスのテストに観測値の 15% を使用します。cvpartition を使用してデータを分割します。
rng("default") % For reproducibility of the partition c = cvpartition(creditrating.Rating,"Holdout",0.15); trainingIndices = training(c); % Indices for the training set testIndices = test(c); % Indices for the test set creditTrain = creditrating(trainingIndices,:); creditTest = creditrating(testIndices,:);
特徴選択の実行
学習データを fitcauto に渡す前に、関数 fscchi2 を使用して重要な予測子を見つけます。関数 bar を使用して、予測子スコアを可視化します。Inf となるスコアがあっても、bar は Inf 値を破棄するため、最初に有限のスコアをプロットしてから、Inf スコアを有限値として別の色でプロットします。
[idx,scores] = fscchi2(creditTrain,"Rating"); bar(scores(idx)) % Represents finite scores hold on veryImportant = isinf(scores); finiteMax = max(scores(~veryImportant)); bar(finiteMax*veryImportant(idx)) % Represents Inf scores hold off xticklabels(strrep(creditTrain.Properties.VariableNames(idx),"_","\_")) xtickangle(45) legend(["Finite Scores","Inf Scores"])
予測子 Industry のスコアは低く、0.05 より大きい p 値に相当します。これは、Industry が重要な特徴ではない可能性があることを示しています。学習データ セットとテスト データ セットから、特徴量 Industry を削除します。
creditTrain = removevars(creditTrain,'Industry'); creditTest = removevars(creditTest,'Industry');
fitcauto の実行
学習データを fitcauto に渡します。この関数は、ベイズ最適化を使用してモデルとそのハイパーパラメーター値を選択し、パフォーマンスが最大となることが期待される学習済みモデル Mdl を返します。使用可能なすべての学習器のタイプを試すよう指定し、最適化を並列実行するよう指定します (Parallel Computing Toolbox™ が必要)。ベイズ最適化の詳細が格納された 2 番目の出力 Results を返します。
このプロセスにいくらか時間がかかることを見込んでください。既定の設定では、fitcauto は、最適化のプロット、および最適化の結果の反復表示を提供します。これらの結果を解釈する方法の詳細については、Verbose の表示を参照してください。
options = struct("UseParallel",true); [Mdl,Results] = fitcauto(creditTrain,"Rating", ... "Learners","all","HyperparameterOptimizationOptions",options);
Starting parallel pool (parpool) using the 'Processes' profile ... 10-May-2024 11:22:29: Job Queued. Waiting for parallel pool job with ID 2 to start ... Connected to parallel pool with 6 workers. Copying objective function to workers... Done copying objective function to workers. Learner types to explore: discr, ensemble, kernel, knn, linear, nb, net, svm, tree Total iterations (MaxObjectiveEvaluations): 270 Total time (MaxTime): Inf |=======================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | validation loss | | | |=======================================================================================================================================================| | 1 | 6 | Best | 0.29704 | 3.5043 | 0.29704 | 0.29704 | tree | MinLeafSize: 9 | | 2 | 6 | Accept | 0.46844 | 3.6329 | 0.29704 | 0.29704 | knn | NumNeighbors: 2 | | | | | | | | | | Distance: cosine | | | | | | | | | | Standardize: false | | 3 | 6 | Accept | 0.47383 | 6.097 | 0.29704 | 0.29704 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.65536 | | | | | | | | | | KernelScale: 3.7007 | | | | | | | | | | Standardize: false | | 4 | 5 | Best | 0.26383 | 6.3788 | 0.26383 | 0.28023 | net | Activations: none | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.00069303 | | | | | | | | | | LayerSizes: 4 | | 5 | 5 | Accept | 0.26533 | 1.2671 | 0.26383 | 0.28023 | tree | MinLeafSize: 48 | | 6 | 5 | Accept | 0.26742 | 0.64789 | 0.26383 | 0.28023 | knn | NumNeighbors: 197 | | | | | | | | | | Distance: cityblock | | | | | | | | | | Standardize: false | | 7 | 5 | Accept | 0.32994 | 0.29613 | 0.26383 | 0.29963 | tree | MinLeafSize: 3 | | 8 | 6 | Accept | 0.68202 | 1.3718 | 0.26383 | 0.29963 | knn | NumNeighbors: 165 | | | | | | | | | | Distance: jaccard | | | | | | | | | | Standardize: false | | 9 | 6 | Accept | 0.68202 | 1.345 | 0.26383 | 0.29963 | knn | NumNeighbors: 165 | | | | | | | | | | Distance: jaccard | | | | | | | | | | Standardize: false | | 10 | 6 | Accept | 0.28059 | 0.79392 | 0.26383 | 0.28059 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 11 | 6 | Accept | 0.73886 | 6.3436 | 0.26383 | 0.28059 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 0.073664 | | | | | | | | | | Lambda: 0.008931 | | | | | | | | | | Standardize: true | | 12 | 6 | Accept | 0.74185 | 4.3986 | 0.26383 | 0.28059 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 0.0016013 | | | | | | | | | | Lambda: 0.043107 | | | | | | | | | | Standardize: false | | 13 | 6 | Accept | 0.28627 | 0.79497 | 0.26383 | 0.28059 | tree | MinLeafSize: 10 | | 14 | 6 | Best | 0.24529 | 2.6341 | 0.24529 | 0.24529 | linear | Coding: onevsone | | | | | | | | | | Lambda: 8.6988e-06 | | | | | | | | | | Learner: svm | | 15 | 6 | Best | 0.242 | 16.312 | 0.242 | 0.24529 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 1.705e-06 | | | | | | | | | | LayerSizes: 2 | | 16 | 5 | Accept | 0.3799 | 0.186 | 0.242 | 0.24529 | tree | MinLeafSize: 492 | | 17 | 5 | Accept | 0.27281 | 33.527 | 0.242 | 0.24529 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 3.3048e-07 | | | | | | | | | | LayerSizes: 34 | | 18 | 6 | Accept | 0.63865 | 0.39589 | 0.242 | 0.24529 | knn | NumNeighbors: 31 | | | | | | | | | | Distance: hamming | | | | | | | | | | Standardize: false | | 19 | 6 | Accept | 0.63865 | 1.7736 | 0.242 | 0.24529 | knn | NumNeighbors: 31 | | | | | | | | | | Distance: hamming | | | | | | | | | | Standardize: false | | 20 | 6 | Accept | 0.74185 | 1.0453 | 0.242 | 0.24529 | discr | Delta: 217.25 | | | | | | | | | | Gamma: 0.30834 | |=======================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | validation loss | | | |=======================================================================================================================================================| | 21 | 4 | Accept | 0.28627 | 29.738 | 0.242 | 0.24529 | ensemble | Coding: onevsone | | | | | | | | | | Method: GentleBoost | | | | | | | | | | NumLearningCycles: 36 | | | | | | | | | | LearnRate: 0.13026 | | | | | | | | | | MinLeafSize: 2 | | 22 | 4 | Accept | 0.24589 | 5.2013 | 0.242 | 0.24529 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 0.9881 | | | | | | | | | | Lambda: 3.9237e-05 | | | | | | | | | | Standardize: false | | 23 | 4 | Accept | 0.42746 | 0.15749 | 0.242 | 0.24529 | discr | Delta: 0.0016931 | | | | | | | | | | Gamma: 0.85944 | | 24 | 6 | Accept | 0.47742 | 0.94715 | 0.242 | 0.25846 | linear | Coding: onevsall | | | | | | | | | | Lambda: 2.4303e-06 | | | | | | | | | | Learner: svm | | 25 | 6 | Accept | 0.47742 | 1.748 | 0.242 | 0.25846 | linear | Coding: onevsall | | | | | | | | | | Lambda: 2.4303e-06 | | | | | | | | | | Learner: svm | | 26 | 5 | Accept | 0.47742 | 2.6179 | 0.242 | 0.28059 | linear | Coding: onevsall | | | | | | | | | | Lambda: 2.4303e-06 | | | | | | | | | | Learner: svm | | 27 | 5 | Accept | 0.74185 | 0.58088 | 0.242 | 0.28059 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.20318 | | | | | | | | | | LayerSizes: 4 | | 28 | 5 | Accept | 0.67424 | 5.7245 | 0.242 | 0.28059 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 4.9742 | | | | | | | | | | Lambda: 0.12271 | | | | | | | | | | Standardize: false | | 29 | 5 | Accept | 0.24379 | 2.0796 | 0.242 | 0.28059 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 0.18363 | | | | | | | | | | KernelScale: 0.068442 | | | | | | | | | | Standardize: false | | 30 | 6 | Accept | 0.56087 | 2.7403 | 0.242 | 0.28059 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.020179 | | | | | | | | | | KernelScale: 45.159 | | | | | | | | | | Standardize: true | | 31 | 6 | Accept | 0.24708 | 1.9624 | 0.242 | 0.28059 | linear | Coding: onevsone | | | | | | | | | | Lambda: 0.0025611 | | | | | | | | | | Learner: logistic | | 32 | 6 | Accept | 0.29704 | 0.20854 | 0.242 | 0.28059 | tree | MinLeafSize: 9 | | 33 | 6 | Accept | 0.74185 | 47.894 | 0.242 | 0.28059 | ensemble | Coding: onevsone | | | | | | | | | | Method: LogitBoost | | | | | | | | | | NumLearningCycles: 77 | | | | | | | | | | LearnRate: 0.0054224 | | | | | | | | | | MinLeafSize: 1430 | | 34 | 5 | Accept | 0.74185 | 4.6386 | 0.242 | 0.28059 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 0.0072048 | | | | | | | | | | Lambda: 0.028163 | | | | | | | | | | Standardize: false | | 35 | 5 | Accept | 0.55639 | 0.65802 | 0.242 | 0.28059 | discr | Delta: 1.0613 | | | | | | | | | | Gamma: 0.24003 | | 36 | 4 | Accept | 0.24379 | 7.4023 | 0.242 | 0.28059 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 1.7808 | | | | | | | | | | Lambda: 3.2594e-05 | | | | | | | | | | Standardize: false | | 37 | 4 | Accept | 0.24678 | 7.2124 | 0.242 | 0.28059 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 1.7808 | | | | | | | | | | Lambda: 3.2594e-05 | | | | | | | | | | Standardize: false | | 38 | 3 | Accept | 0.67155 | 1.7195 | 0.242 | 0.28059 | knn | NumNeighbors: 1445 | | | | | | | | | | Distance: spearman | | | | | | | | | | Standardize: true | | 39 | 3 | Accept | 0.46455 | 0.60315 | 0.242 | 0.28059 | linear | Coding: onevsall | | | | | | | | | | Lambda: 0.3312 | | | | | | | | | | Learner: logistic | | 40 | 6 | Accept | 0.28059 | 0.14968 | 0.242 | 0.28059 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | |=======================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | validation loss | | | |=======================================================================================================================================================| | 41 | 5 | Accept | 0.61113 | 0.31292 | 0.242 | 0.28059 | tree | MinLeafSize: 881 | | 42 | 5 | Accept | 0.28059 | 0.75648 | 0.242 | 0.28059 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 43 | 5 | Accept | 0.29494 | 1.4567 | 0.242 | 0.28059 | linear | Coding: onevsone | | | | | | | | | | Lambda: 0.03378 | | | | | | | | | | Learner: svm | | 44 | 5 | Accept | 0.43823 | 0.74458 | 0.242 | 0.28059 | linear | Coding: onevsall | | | | | | | | | | Lambda: 0.0084424 | | | | | | | | | | Learner: logistic | | 45 | 5 | Accept | 0.71702 | 0.49019 | 0.242 | 0.28059 | knn | NumNeighbors: 457 | | | | | | | | | | Distance: jaccard | | | | | | | | | | Standardize: true | | 46 | 6 | Accept | 0.24738 | 2.2676 | 0.242 | 0.28059 | linear | Coding: onevsone | | | | | | | | | | Lambda: 1.8074e-05 | | | | | | | | | | Learner: logistic | | 47 | 4 | Accept | 0.45648 | 8.8659 | 0.242 | 0.26483 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.0075021 | | | | | | | | | | KernelScale: 0.023323 | | | | | | | | | | Standardize: true | | 48 | 4 | Accept | 0.47203 | 4.1096 | 0.242 | 0.26483 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 1.4724 | | | | | | | | | | Lambda: 0.063523 | | | | | | | | | | Standardize: true | | 49 | 4 | Accept | 0.24738 | 3.0612 | 0.242 | 0.26483 | linear | Coding: onevsone | | | | | | | | | | Lambda: 1.8074e-05 | | | | | | | | | | Learner: logistic | | 50 | 4 | Accept | 0.24768 | 1.7307 | 0.242 | 0.26483 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 0.013982 | | | | | | | | | | KernelScale: 0.1266 | | | | | | | | | | Standardize: true | | 51 | 6 | Accept | 0.74185 | 2.788 | 0.242 | 0.26483 | nb | DistributionNames: kernel | | | | | | | | | | Width: 417.63 | | | | | | | | | | Standardize: false | | 52 | 6 | Accept | 0.42716 | 0.11455 | 0.242 | 0.26483 | discr | Delta: 0.021092 | | | | | | | | | | Gamma: 0.256 | | 53 | 6 | Accept | 0.24349 | 76.191 | 0.242 | 0.26483 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 1.3715e-05 | | | | | | | | | | LayerSizes: 122 | | 54 | 6 | Accept | 0.43314 | 0.9632 | 0.242 | 0.26483 | discr | Delta: 9.1157e-06 | | | | | | | | | | Gamma: 0.53332 | | 55 | 6 | Accept | 0.74185 | 4.3906 | 0.242 | 0.26483 | nb | DistributionNames: kernel | | | | | | | | | | Width: 18.623 | | | | | | | | | | Standardize: true | | 56 | 6 | Accept | 0.45917 | 8.4506 | 0.242 | 0.26483 | net | Activations: relu | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.007474 | | | | | | | | | | LayerSizes: [ 1 20 ] | | 57 | 6 | Accept | 0.25456 | 0.31225 | 0.242 | 0.26483 | knn | NumNeighbors: 21 | | | | | | | | | | Distance: euclidean | | | | | | | | | | Standardize: false | | 58 | 6 | Accept | 0.24319 | 20.176 | 0.242 | 0.26454 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.5101e-05 | | | | | | | | | | LayerSizes: 17 | | 59 | 6 | Accept | 0.38139 | 1.3557 | 0.242 | 0.26454 | knn | NumNeighbors: 112 | | | | | | | | | | Distance: cosine | | | | | | | | | | Standardize: true | | 60 | 6 | Accept | 0.48699 | 2.5673 | 0.242 | 0.26188 | linear | Coding: onevsall | | | | | | | | | | Lambda: 0.85779 | | | | | | | | | | Learner: svm | |=======================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | validation loss | | | |=======================================================================================================================================================| | 61 | 6 | Accept | 0.24439 | 1.8066 | 0.242 | 0.25682 | linear | Coding: onevsone | | | | | | | | | | Lambda: 1.7606e-08 | | | | | | | | | | Learner: svm | | 62 | 6 | Accept | 0.54442 | 1.1095 | 0.242 | 0.25682 | knn | NumNeighbors: 1256 | | | | | | | | | | Distance: cosine | | | | | | | | | | Standardize: true | | 63 | 6 | Accept | 0.28059 | 0.25665 | 0.242 | 0.25682 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 64 | 6 | Accept | 0.78283 | 9.2719 | 0.242 | 0.25682 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 0.0084622 | | | | | | | | | | Lambda: 4.3995e-05 | | | | | | | | | | Standardize: true | | 65 | 6 | Accept | 0.43284 | 9.9414 | 0.242 | 0.25682 | ensemble | Method: Bag | | | | | | | | | | NumLearningCycles: 171 | | | | | | | | | | LearnRate: NaN | | | | | | | | | | MinLeafSize: 583 | | 66 | 6 | Accept | 0.56985 | 0.16211 | 0.242 | 0.25682 | discr | Delta: 1.0473 | | | | | | | | | | Gamma: 0.72754 | | 67 | 6 | Accept | 0.24559 | 17.448 | 0.242 | 0.25682 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 0.30375 | | | | | | | | | | KernelScale: 0.094485 | | | | | | | | | | Standardize: true | | 68 | 6 | Accept | 0.46246 | 2.2288 | 0.242 | 0.25682 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 15.954 | | | | | | | | | | KernelScale: 3.1701 | | | | | | | | | | Standardize: false | | 69 | 6 | Accept | 0.74065 | 4.0667 | 0.242 | 0.25682 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 0.027004 | | | | | | | | | | Lambda: 0.010714 | | | | | | | | | | Standardize: false | | 70 | 6 | Accept | 0.3117 | 117.25 | 0.242 | 0.25682 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 5.5502e-07 | | | | | | | | | | LayerSizes: 141 | | 71 | 6 | Accept | 0.37571 | 0.32565 | 0.242 | 0.25682 | tree | MinLeafSize: 469 | | 72 | 6 | Accept | 0.28836 | 2.5757 | 0.242 | 0.25682 | nb | DistributionNames: kernel | | | | | | | | | | Width: 0.049199 | | | | | | | | | | Standardize: true | | 73 | 6 | Accept | 0.71941 | 0.57072 | 0.242 | 0.25682 | knn | NumNeighbors: 681 | | | | | | | | | | Distance: jaccard | | | | | | | | | | Standardize: true | | 74 | 6 | Accept | 0.54472 | 0.57641 | 0.242 | 0.25682 | discr | Delta: 0.89876 | | | | | | | | | | Gamma: 0.34215 | | 75 | 6 | Accept | 0.53814 | 0.95437 | 0.242 | 0.25682 | ensemble | Method: Bag | | | | | | | | | | NumLearningCycles: 16 | | | | | | | | | | LearnRate: NaN | | | | | | | | | | MinLeafSize: 1027 | | 76 | 6 | Accept | 0.5175 | 3.6894 | 0.242 | 0.25682 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.0022428 | | | | | | | | | | LayerSizes: [ 65 2 ] | | 77 | 6 | Accept | 0.28896 | 1.2373 | 0.242 | 0.25682 | nb | DistributionNames: kernel | | | | | | | | | | Width: 0.026834 | | | | | | | | | | Standardize: true | | 78 | 6 | Accept | 0.74005 | 2.4 | 0.242 | 0.25682 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 408.15 | | | | | | | | | | Lambda: 0.00054653 | | | | | | | | | | Standardize: false | | 79 | 6 | Accept | 0.71463 | 2.3243 | 0.242 | 0.25682 | nb | DistributionNames: kernel | | | | | | | | | | Width: 5.7434 | | | | | | | | | | Standardize: false | | 80 | 6 | Accept | 0.31768 | 1.4475 | 0.242 | 0.25281 | linear | Coding: onevsone | | | | | | | | | | Lambda: 0.14117 | | | | | | | | | | Learner: svm | |=======================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | validation loss | | | |=======================================================================================================================================================| | 81 | 6 | Accept | 0.42926 | 0.10631 | 0.242 | 0.25281 | discr | Delta: 0.0056255 | | | | | | | | | | Gamma: 0.29449 | | 82 | 6 | Accept | 0.74185 | 3.7651 | 0.242 | 0.25281 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 0.0021877 | | | | | | | | | | Lambda: 0.12375 | | | | | | | | | | Standardize: true | | 83 | 6 | Accept | 0.27879 | 196.9 | 0.242 | 0.25281 | ensemble | Coding: onevsone | | | | | | | | | | Method: GentleBoost | | | | | | | | | | NumLearningCycles: 306 | | | | | | | | | | LearnRate: 0.0028268 | | | | | | | | | | MinLeafSize: 5 | | 84 | 6 | Accept | 0.69518 | 0.10299 | 0.242 | 0.25281 | discr | Delta: 2.2186 | | | | | | | | | | Gamma: 0.48692 | | 85 | 6 | Accept | 0.46515 | 0.10863 | 0.242 | 0.25281 | tree | MinLeafSize: 666 | | 86 | 6 | Accept | 0.25845 | 11.362 | 0.242 | 0.25281 | ensemble | Method: AdaBoostM2 | | | | | | | | | | NumLearningCycles: 232 | | | | | | | | | | LearnRate: 0.057232 | | | | | | | | | | MinLeafSize: 2 | | 87 | 6 | Accept | 0.4819 | 0.23471 | 0.242 | 0.25281 | knn | NumNeighbors: 18 | | | | | | | | | | Distance: correlation | | | | | | | | | | Standardize: false | | 88 | 6 | Accept | 0.59497 | 0.64377 | 0.242 | 0.25281 | knn | NumNeighbors: 52 | | | | | | | | | | Distance: spearman | | | | | | | | | | Standardize: true | | 89 | 6 | Accept | 0.25965 | 7.2545 | 0.242 | 0.25281 | ensemble | Method: Bag | | | | | | | | | | NumLearningCycles: 66 | | | | | | | | | | LearnRate: NaN | | | | | | | | | | MinLeafSize: 1 | | 90 | 6 | Accept | 0.74185 | 1.158 | 0.242 | 0.25281 | net | Activations: tanh | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 8.4228 | | | | | | | | | | LayerSizes: [ 75 2 87 ] | | 91 | 6 | Accept | 0.26653 | 1.1482 | 0.242 | 0.25281 | ensemble | Method: AdaBoostM2 | | | | | | | | | | NumLearningCycles: 15 | | | | | | | | | | LearnRate: 0.11472 | | | | | | | | | | MinLeafSize: 39 | | 92 | 6 | Accept | 0.45049 | 5.8162 | 0.242 | 0.25281 | ensemble | Method: Bag | | | | | | | | | | NumLearningCycles: 127 | | | | | | | | | | LearnRate: NaN | | | | | | | | | | MinLeafSize: 830 | | 93 | 6 | Accept | 0.29225 | 5.0049 | 0.242 | 0.25281 | ensemble | Method: AdaBoostM2 | | | | | | | | | | NumLearningCycles: 110 | | | | | | | | | | LearnRate: 0.0018187 | | | | | | | | | | MinLeafSize: 15 | | 94 | 6 | Accept | 0.39097 | 0.88194 | 0.242 | 0.25281 | nb | DistributionNames: kernel | | | | | | | | | | Width: 0.0020466 | | | | | | | | | | Standardize: true | | 95 | 6 | Accept | 0.28059 | 0.16411 | 0.242 | 0.25281 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 96 | 6 | Accept | 0.74185 | 0.75488 | 0.242 | 0.25281 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.010761 | | | | | | | | | | LayerSizes: [ 18 8 17 ] | | 97 | 6 | Accept | 0.28059 | 0.11286 | 0.242 | 0.25281 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 98 | 6 | Accept | 0.28059 | 0.081407 | 0.242 | 0.25281 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 99 | 6 | Accept | 0.28627 | 0.21186 | 0.242 | 0.25281 | tree | MinLeafSize: 10 | | 100 | 6 | Accept | 0.74185 | 2.2671 | 0.242 | 0.25281 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 0.0013533 | | | | | | | | | | Lambda: 0.031901 | | | | | | | | | | Standardize: false | |=======================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | validation loss | | | |=======================================================================================================================================================| | 101 | 6 | Accept | 0.61322 | 1.0461 | 0.242 | 0.25281 | knn | NumNeighbors: 435 | | | | | | | | | | Distance: spearman | | | | | | | | | | Standardize: true | | 102 | 6 | Accept | 0.28059 | 0.12504 | 0.242 | 0.25281 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 103 | 6 | Accept | 0.28059 | 0.12053 | 0.242 | 0.25281 | nb | DistributionNames: normal | | | | | | | | | | Width: NaN | | | | | | | | | | Standardize: - | | 104 | 6 | Accept | 0.45169 | 0.81334 | 0.242 | 0.25215 | linear | Coding: onevsall | | | | | | | | | | Lambda: 0.017986 | | | | | | | | | | Learner: svm | | 105 | 6 | Accept | 0.54831 | 0.25986 | 0.242 | 0.25215 | knn | NumNeighbors: 50 | | | | | | | | | | Distance: correlation | | | | | | | | | | Standardize: true | | 106 | 6 | Accept | 0.4496 | 324.95 | 0.242 | 0.25215 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 0.0012647 | | | | | | | | | | KernelScale: 0.0017986 | | | | | | | | | | Standardize: true | | 107 | 6 | Accept | 0.3781 | 1.8041 | 0.242 | 0.25434 | linear | Coding: onevsall | | | | | | | | | | Lambda: 4.5023e-06 | | | | | | | | | | Learner: logistic | | 108 | 6 | Accept | 0.44451 | 0.70126 | 0.242 | 0.25462 | linear | Coding: onevsall | | | | | | | | | | Lambda: 0.12817 | | | | | | | | | | Learner: logistic | | 109 | 6 | Accept | 0.26025 | 23.436 | 0.242 | 0.25462 | ensemble | Coding: onevsall | | | | | | | | | | Method: AdaBoostM1 | | | | | | | | | | NumLearningCycles: 72 | | | | | | | | | | LearnRate: 0.015914 | | | | | | | | | | MinLeafSize: 134 | | 110 | 6 | Accept | 0.74185 | 0.55177 | 0.242 | 0.25462 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.16788 | | | | | | | | | | LayerSizes: [ 1 12 56 ] | | 111 | 6 | Accept | 0.4143 | 10.391 | 0.242 | 0.25462 | ensemble | Method: Bag | | | | | | | | | | NumLearningCycles: 175 | | | | | | | | | | LearnRate: NaN | | | | | | | | | | MinLeafSize: 563 | | 112 | 6 | Accept | 0.56506 | 0.11833 | 0.242 | 0.25462 | discr | Delta: 1.3228 | | | | | | | | | | Gamma: 0.34287 | | 113 | 6 | Accept | 0.64433 | 0.30407 | 0.242 | 0.25462 | knn | NumNeighbors: 26 | | | | | | | | | | Distance: hamming | | | | | | | | | | Standardize: false | | 114 | 6 | Accept | 0.27789 | 1.4143 | 0.242 | 0.25462 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 0.12591 | | | | | | | | | | KernelScale: 2.1296 | | | | | | | | | | Standardize: true | | 115 | 6 | Accept | 0.47113 | 0.16139 | 0.242 | 0.25462 | tree | MinLeafSize: 674 | | 116 | 6 | Accept | 0.74185 | 1.991 | 0.242 | 0.25462 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 525.99 | | | | | | | | | | Lambda: 0.024827 | | | | | | | | | | Standardize: true | | 117 | 6 | Accept | 0.45319 | 15.81 | 0.242 | 0.25462 | svm | Coding: onevsall | | | | | | | | | | BoxConstraint: 29.239 | | | | | | | | | | KernelScale: 0.37967 | | | | | | | | | | Standardize: false | | 118 | 6 | Accept | 0.47861 | 498.57 | 0.242 | 0.25462 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 127.44 | | | | | | | | | | KernelScale: 0.0098561 | | | | | | | | | | Standardize: true | | 119 | 6 | Accept | 0.32785 | 0.23811 | 0.242 | 0.25462 | tree | MinLeafSize: 4 | | 120 | 6 | Accept | 0.42985 | 0.1125 | 0.242 | 0.25462 | discr | Delta: 0.00023581 | | | | | | | | | | Gamma: 0.74169 | |=======================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | validation loss | | | |=======================================================================================================================================================| | 121 | 6 | Accept | 0.74185 | 0.28212 | 0.242 | 0.25462 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 2.7198 | | | | | | | | | | LayerSizes: 3 | | 122 | 6 | Accept | 0.29016 | 12.52 | 0.242 | 0.25462 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 0.24974 | | | | | | | | | | Lambda: 1.6898e-06 | | | | | | | | | | Standardize: false | | 123 | 6 | Accept | 0.61681 | 0.082854 | 0.242 | 0.25462 | tree | MinLeafSize: 1037 | | 124 | 6 | Accept | 0.28687 | 35.279 | 0.242 | 0.25462 | ensemble | Coding: onevsall | | | | | | | | | | Method: LogitBoost | | | | | | | | | | NumLearningCycles: 68 | | | | | | | | | | LearnRate: 0.99683 | | | | | | | | | | MinLeafSize: 9 | | 125 | 6 | Accept | 0.3793 | 0.16321 | 0.242 | 0.25462 | tree | MinLeafSize: 430 | | 126 | 6 | Accept | 0.2426 | 34.709 | 0.242 | 0.25462 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.0002e-07 | | | | | | | | | | LayerSizes: [ 2 128 ] | | 127 | 6 | Accept | 0.42656 | 0.12103 | 0.242 | 0.25462 | discr | Delta: 1.8328e-06 | | | | | | | | | | Gamma: 0.14399 | | 128 | 6 | Accept | 0.42537 | 0.089666 | 0.242 | 0.25462 | discr | Delta: 1.73e-05 | | | | | | | | | | Gamma: 0.15184 | | 129 | 6 | Accept | 0.24349 | 1.5715 | 0.242 | 0.25462 | svm | Coding: onevsone | | | | | | | | | | BoxConstraint: 6.9807 | | | | | | | | | | KernelScale: 0.60748 | | | | | | | | | | Standardize: false | | 130 | 6 | Accept | 0.79539 | 3.6791 | 0.242 | 0.25462 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 0.018788 | | | | | | | | | | Lambda: 1.0474e-05 | | | | | | | | | | Standardize: false | | 131 | 6 | Accept | 0.26862 | 11.273 | 0.242 | 0.25462 | ensemble | Method: AdaBoostM2 | | | | | | | | | | NumLearningCycles: 204 | | | | | | | | | | LearnRate: 0.0057787 | | | | | | | | | | MinLeafSize: 5 | | 132 | 6 | Accept | 0.48041 | 5.0926 | 0.242 | 0.25462 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 0.23411 | | | | | | | | | | Lambda: 3.608e-05 | | | | | | | | | | Standardize: true | | 133 | 6 | Accept | 0.30242 | 5.5709 | 0.242 | 0.25462 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 126.67 | | | | | | | | | | Lambda: 3.1052e-06 | | | | | | | | | | Standardize: false | | 134 | 6 | Accept | 0.24678 | 2.7052 | 0.242 | 0.25172 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.00059683 | | | | | | | | | | LayerSizes: 3 | | 135 | 6 | Accept | 0.73587 | 4.6448 | 0.242 | 0.25172 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 736.62 | | | | | | | | | | Lambda: 3.7849e-05 | | | | | | | | | | Standardize: true | | 136 | 6 | Accept | 0.4487 | 3.0531 | 0.242 | 0.25172 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 752.25 | | | | | | | | | | Lambda: 5.4668e-07 | | | | | | | | | | Standardize: false | | 137 | 6 | Best | 0.2414 | 11.239 | 0.2414 | 0.25462 | net | Activations: none | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.00046679 | | | | | | | | | | LayerSizes: [ 5 91 4 ] | | 138 | 6 | Accept | 0.25396 | 35.835 | 0.2414 | 0.25241 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 2.1901e-06 | | | | | | | | | | LayerSizes: [ 11 2 37 ] | | 139 | 6 | Accept | 0.45707 | 1.0704 | 0.2414 | 0.25462 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.00118 | | | | | | | | | | LayerSizes: [ 2 2 ] | | 140 | 6 | Accept | 0.2414 | 15.673 | 0.2414 | 0.24974 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 4.1041e-05 | | | | | | | | | | LayerSizes: [ 2 5 16 ] | |=======================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | validation loss | | | |=======================================================================================================================================================| | 141 | 6 | Accept | 0.24589 | 5.5905 | 0.2414 | 0.24974 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 6.9075 | | | | | | | | | | Lambda: 2.345e-05 | | | | | | | | | | Standardize: false | | 142 | 6 | Accept | 0.27311 | 6.419 | 0.2414 | 0.24974 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 1.1733 | | | | | | | | | | Lambda: 3.215e-05 | | | | | | | | | | Standardize: true | | 143 | 6 | Accept | 0.24649 | 5.5317 | 0.2414 | 0.24974 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 3.828 | | | | | | | | | | Lambda: 3.5924e-05 | | | | | | | | | | Standardize: false | | 144 | 6 | Accept | 0.42806 | 8.2774 | 0.2414 | 0.24974 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 0.093493 | | | | | | | | | | Lambda: 3.1039e-05 | | | | | | | | | | Standardize: false | | 145 | 6 | Accept | 0.37063 | 5.2086 | 0.2414 | 0.24974 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 102.4 | | | | | | | | | | Lambda: 3.5109e-05 | | | | | | | | | | Standardize: true | | 146 | 6 | Accept | 0.26234 | 5.3734 | 0.2414 | 0.24974 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 17.801 | | | | | | | | | | Lambda: 3.648e-05 | | | | | | | | | | Standardize: false | | 147 | 6 | Accept | 0.24888 | 5.7508 | 0.2414 | 0.24974 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 2.1707 | | | | | | | | | | Lambda: 1.0135e-05 | | | | | | | | | | Standardize: false | | 148 | 6 | Accept | 0.29674 | 4.1202 | 0.2414 | 0.24974 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 6.3863 | | | | | | | | | | Lambda: 3.9393e-05 | | | | | | | | | | Standardize: false | | 149 | 6 | Accept | 0.26413 | 7.9572 | 0.2414 | 0.24974 | ensemble | Coding: onevsall | | | | | | | | | | Method: LogitBoost | | | | | | | | | | NumLearningCycles: 22 | | | | | | | | | | LearnRate: 0.052662 | | | | | | | | | | MinLeafSize: 82 | | 150 | 6 | Accept | 0.29734 | 4.2389 | 0.2414 | 0.24974 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 6.1493 | | | | | | | | | | Lambda: 3.7581e-05 | | | | | | | | | | Standardize: false | | 151 | 6 | Accept | 0.29614 | 2.433 | 0.2414 | 0.24974 | ensemble | Method: AdaBoostM2 | | | | | | | | | | NumLearningCycles: 45 | | | | | | | | | | LearnRate: 0.037037 | | | | | | | | | | MinLeafSize: 215 | | 152 | 6 | Accept | 0.25187 | 6.677 | 0.2414 | 0.24974 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 4.4623 | | | | | | | | | | Lambda: 3.9302e-05 | | | | | | | | | | Standardize: true | | 153 | 6 | Accept | 0.24499 | 5.6011 | 0.2414 | 0.24974 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 3.6605 | | | | | | | | | | Lambda: 4.0695e-05 | | | | | | | | | | Standardize: false | | 154 | 6 | Best | 0.2399 | 10.914 | 0.2399 | 0.24458 | net | Activations: none | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.00087823 | | | | | | | | | | LayerSizes: [ 3 10 135 ] | | 155 | 6 | Accept | 0.26533 | 4.3657 | 0.2399 | 0.24458 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 3.1699 | | | | | | | | | | Lambda: 4.2302e-05 | | | | | | | | | | Standardize: false | | 156 | 6 | Accept | 0.25037 | 4.4619 | 0.2399 | 0.24458 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 1.4452 | | | | | | | | | | Lambda: 4.3079e-05 | | | | | | | | | | Standardize: false | | 157 | 6 | Accept | 0.24499 | 5.4242 | 0.2399 | 0.24458 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 3.7694 | | | | | | | | | | Lambda: 4.3534e-05 | | | | | | | | | | Standardize: false | | 158 | 6 | Accept | 0.26563 | 4.2206 | 0.2399 | 0.24458 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 3.1282 | | | | | | | | | | Lambda: 4.3446e-05 | | | | | | | | | | Standardize: false | | 159 | 6 | Accept | 0.24499 | 5.0307 | 0.2399 | 0.24458 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 4.0207 | | | | | | | | | | Lambda: 7.2479e-05 | | | | | | | | | | Standardize: false | | 160 | 6 | Accept | 0.24379 | 5.4303 | 0.2399 | 0.24458 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 4.6559 | | | | | | | | | | Lambda: 4.0574e-05 | | | | | | | | | | Standardize: false | |=======================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | validation loss | | | |=======================================================================================================================================================| | 161 | 6 | Accept | 0.24738 | 4.2359 | 0.2399 | 0.24458 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 1.8848 | | | | | | | | | | Lambda: 4.73e-05 | | | | | | | | | | Standardize: false | | 162 | 6 | Accept | 0.24738 | 5.267 | 0.2399 | 0.24458 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 0.80973 | | | | | | | | | | Lambda: 0.00016926 | | | | | | | | | | Standardize: false | | 163 | 6 | Accept | 0.24858 | 3.8912 | 0.2399 | 0.24458 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 2.002 | | | | | | | | | | Lambda: 6.9602e-05 | | | | | | | | | | Standardize: false | | 164 | 6 | Accept | 0.25396 | 3.3102 | 0.2399 | 0.24458 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 1.7161 | | | | | | | | | | Lambda: 0.00024265 | | | | | | | | | | Standardize: false | | 165 | 6 | Accept | 0.32486 | 3.3696 | 0.2399 | 0.24458 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 5.1868 | | | | | | | | | | Lambda: 0.00030745 | | | | | | | | | | Standardize: false | | 166 | 6 | Accept | 0.26593 | 6.9799 | 0.2399 | 0.24458 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 1.8249 | | | | | | | | | | Lambda: 5.4968e-05 | | | | | | | | | | Standardize: true | | 167 | 6 | Accept | 0.24858 | 4.795 | 0.2399 | 0.24458 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 1.0636 | | | | | | | | | | Lambda: 0.00040238 | | | | | | | | | | Standardize: false | | 168 | 6 | Accept | 0.25755 | 6.1694 | 0.2399 | 0.24458 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 2.245 | | | | | | | | | | Lambda: 6.6618e-05 | | | | | | | | | | Standardize: true | | 169 | 6 | Accept | 0.26234 | 3.0205 | 0.2399 | 0.24458 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 1.6204 | | | | | | | | | | Lambda: 0.00044561 | | | | | | | | | | Standardize: false | | 170 | 6 | Accept | 0.27161 | 3.0247 | 0.2399 | 0.24458 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 2.0168 | | | | | | | | | | Lambda: 0.00041694 | | | | | | | | | | Standardize: false | | 171 | 6 | Accept | 0.25217 | 8.6916 | 0.2399 | 0.24458 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 3.9761 | | | | | | | | | | Lambda: 8.9879e-06 | | | | | | | | | | Standardize: true | | 172 | 6 | Accept | 0.26653 | 4.4537 | 0.2399 | 0.24458 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 3.5929 | | | | | | | | | | Lambda: 0.00096967 | | | | | | | | | | Standardize: false | | 173 | 6 | Accept | 0.25666 | 6.5672 | 0.2399 | 0.24458 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 0.40713 | | | | | | | | | | Lambda: 4.6217e-05 | | | | | | | | | | Standardize: false | | 174 | 6 | Accept | 0.33204 | 3.0506 | 0.2399 | 0.24458 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 4.4378 | | | | | | | | | | Lambda: 0.00050652 | | | | | | | | | | Standardize: false | | 175 | 6 | Accept | 0.28866 | 2.9742 | 0.2399 | 0.24458 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 2.3933 | | | | | | | | | | Lambda: 0.00058278 | | | | | | | | | | Standardize: false | | 176 | 6 | Accept | 0.25217 | 4.854 | 0.2399 | 0.24458 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 4.8744 | | | | | | | | | | Lambda: 0.00030213 | | | | | | | | | | Standardize: false | | 177 | 6 | Accept | 0.37212 | 2.4693 | 0.2399 | 0.24458 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 3.8515 | | | | | | | | | | Lambda: 0.0029253 | | | | | | | | | | Standardize: false | | 178 | 6 | Accept | 0.24888 | 2.9753 | 0.2399 | 0.24458 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 1.0426 | | | | | | | | | | Lambda: 0.00064459 | | | | | | | | | | Standardize: false | | 179 | 6 | Accept | 0.25695 | 5.1847 | 0.2399 | 0.24458 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 4.8233 | | | | | | | | | | Lambda: 0.00023726 | | | | | | | | | | Standardize: true | | 180 | 6 | Accept | 0.26324 | 4.8044 | 0.2399 | 0.24458 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 4.8311 | | | | | | | | | | Lambda: 0.00054379 | | | | | | | | | | Standardize: true | |=======================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | validation loss | | | |=======================================================================================================================================================| | 181 | 6 | Accept | 0.24708 | 4.3993 | 0.2399 | 0.24458 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 0.97694 | | | | | | | | | | Lambda: 0.0030881 | | | | | | | | | | Standardize: false | | 182 | 6 | Accept | 0.25337 | 2.4427 | 0.2399 | 0.24714 | net | Activations: none | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.001292 | | | | | | | | | | LayerSizes: [ 26 4 ] | | 183 | 6 | Accept | 0.39067 | 2.5062 | 0.2399 | 0.24634 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 1.2819 | | | | | | | | | | Lambda: 0.0052145 | | | | | | | | | | Standardize: true | | 184 | 6 | Accept | 0.24678 | 4.7713 | 0.2399 | 0.24774 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 0.87503 | | | | | | | | | | Lambda: 0.0005457 | | | | | | | | | | Standardize: false | | 185 | 6 | Accept | 0.32725 | 3.0177 | 0.2399 | 0.24622 | kernel | Coding: onevsall | | | | | | | | | | KernelScale: 1.1776 | | | | | | | | | | Lambda: 0.00077561 | | | | | | | | | | Standardize: true | | 186 | 6 | Accept | 0.25486 | 5.2826 | 0.2399 | 0.24654 | kernel | Coding: onevsone | | | | | | | | | | KernelScale: 5.1429 | | | | | | | | | | Lambda: 0.00014913 | | | | | | | | | | Standardize: true | | 187 | 6 | Accept | 0.2405 | 13.417 | 0.2399 | 0.24374 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.00098014 | | | | | | | | | | LayerSizes: [ 3 168 6 ] | | 188 | 6 | Accept | 0.26772 | 5.7891 | 0.2399 | 0.24613 | net | Activations: none | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.0016998 | | | | | | | | | | LayerSizes: [ 1 222 ] | | 189 | 6 | Accept | 0.242 | 28.819 | 0.2399 | 0.24346 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.0765e-05 | | | | | | | | | | LayerSizes: [ 2 34 ] | | 190 | 6 | Accept | 0.2399 | 17.431 | 0.2399 | 0.2443 | net | Activations: relu | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.799e-05 | | | | | | | | | | LayerSizes: [ 6 2 13 ] | | 191 | 6 | Accept | 0.64373 | 4.0354 | 0.2399 | 0.24597 | net | Activations: relu | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 9.6616e-06 | | | | | | | | | | LayerSizes: [ 1 1 16 ] | | 192 | 6 | Accept | 0.29195 | 7.1387 | 0.2399 | 0.24504 | net | Activations: none | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.0018314 | | | | | | | | | | LayerSizes: [ 279 1 1 ] | | 193 | 6 | Accept | 0.24738 | 17.111 | 0.2399 | 0.24654 | net | Activations: relu | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 2.7262e-05 | | | | | | | | | | LayerSizes: [ 2 3 12 ] | | 194 | 6 | Accept | 0.25695 | 19.097 | 0.2399 | 0.2451 | net | Activations: tanh | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.0018592 | | | | | | | | | | LayerSizes: [ 5 8 21 ] | | 195 | 6 | Accept | 0.30482 | 7.588 | 0.2399 | 0.24529 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 1.6205e-05 | | | | | | | | | | LayerSizes: [ 3 1 ] | | 196 | 6 | Accept | 0.2426 | 16.05 | 0.2399 | 0.24348 | net | Activations: none | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 4.2692e-05 | | | | | | | | | | LayerSizes: [ 1 61 13 ] | | 197 | 6 | Accept | 0.25606 | 37.701 | 0.2399 | 0.24654 | net | Activations: relu | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 3.4757e-06 | | | | | | | | | | LayerSizes: [ 6 90 4 ] | | 198 | 6 | Accept | 0.2405 | 44.824 | 0.2399 | 0.2427 | net | Activations: relu | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.00071975 | | | | | | | | | | LayerSizes: [ 2 14 102 ] | | 199 | 6 | Accept | 0.26533 | 5.2454 | 0.2399 | 0.24562 | net | Activations: none | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.0031705 | | | | | | | | | | LayerSizes: [ 6 17 145 ] | | 200 | 6 | Accept | 0.28687 | 2.6155 | 0.2399 | 0.24499 | net | Activations: none | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.004398 | | | | | | | | | | LayerSizes: [ 2 69 9 ] | |=======================================================================================================================================================| | Iter | Active | Eval | Validation | Time for training | Observed min | Estimated min | Learner | Hyperparameter: Value | | | workers | result | loss | & validation (sec)| validation loss | validation loss | | | |=======================================================================================================================================================| | 201 | 6 | Accept | 0.24738 | 18.077 | 0.2399 | 0.24507 | net | Activations: relu | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 1.5557e-05 | | | | | | | | | | LayerSizes: [ 3 12 13 ] | | 202 | 6 | Accept | 0.2402 | 10.51 | 0.2399 | 0.24258 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 4.2267e-05 | | | | | | | | | | LayerSizes: [ 1 7 5 ] | | 203 | 6 | Accept | 0.33712 | 16.279 | 0.2399 | 0.24512 | net | Activations: none | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.0066296 | | | | | | | | | | LayerSizes: [ 3 225 228 ] | | 204 | 6 | Accept | 0.74185 | 0.57522 | 0.2399 | 0.24528 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.0022415 | | | | | | | | | | LayerSizes: [ 4 5 2 ] | | 205 | 6 | Accept | 0.2411 | 9.1254 | 0.2399 | 0.24401 | net | Activations: none | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 3.6813e-09 | | | | | | | | | | LayerSizes: [ 1 9 16 ] | | 206 | 6 | Accept | 0.74185 | 1.6902 | 0.2399 | 0.24501 | net | Activations: sigmoid | | | | | | | | | | Standardize: true | | | | | | | | | | Lambda: 0.0018843 | | | | | | | | | | LayerSizes: [ 26 5 52 ] | | 207 | 6 | Accept | 0.74185 | 0.96211 | 0.2399 | 0.24422 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.0010998 | | | | | | | | | | LayerSizes: [ 24 4 6 ] | | 208 | 6 | Accept | 0.24559 | 20.386 | 0.2399 | 0.24513 | net | Activations: tanh | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 3.1073e-05 | | | | | | | | | | LayerSizes: [ 2 2 2 ] | | 209 | 6 | Accept | 0.28418 | 16.286 | 0.2399 | 0.24432 | net | Activations: tanh | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.00041184 | | | | | | | | | | LayerSizes: [ 3 5 2 ] | | 210 | 6 | Accept | 0.29494 | 21.129 | 0.2399 | 0.24208 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | | | | | Lambda: 0.00051161 | | | | | | | | | | LayerSizes: [ 11 29 20 ] | | 211 | 6 | Accept | 0.2411 | 68.677 | 0.2399 | 0.24394 | net | Activations: sigmoid | | | | | | | | | | Standardize: false | | | | | | ...
__________________________________________________________ Optimization completed. Total iterations: 270 Total elapsed time: 2056.3088 seconds Total time for training and validation: 11269.3988 seconds Best observed learner is a net model with: Learner: net Activations: none Standardize: true Lambda: 0.00080617 LayerSizes: [24 293 67] Observed validation loss: 0.23931 Time for training and validation: 35.4452 seconds Best estimated learner (returned model) is a net model with: Learner: net Activations: none Standardize: false Lambda: 0.001083 LayerSizes: [4 18 18] Estimated validation loss: 0.24198 Estimated time for training and validation: 6.818 seconds Documentation for fitcauto display
fitcauto によって返される最終的なモデルが、最適な推定学習器となります。モデルを返す前に、関数は学習データ全体 (creditTrain)、リストされている Learner (またはモデル) のタイプ、および表示されたハイパーパラメーター値を使用して、モデルの再学習を行います。
テスト セットのパフォーマンスの評価
モデル Mdl は、基準 "min-visited-mean" に従った、ベイズ最適化における最適な点に対応します。新しいデータに対するモデルのパフォーマンスを評価するには、モデルで観測された交差検証精度 (cvAccuracy)、およびベイズ最適化に基づき推定されたモデルの全体的なパフォーマンス (estimatedAccuracy) を確認します。
[x,~,iteration] = bestPoint(Results,"Criterion","min-visited-mean"); cvError = Results.ObjectiveTrace(iteration); cvAccuracy = 1 - cvError
cvAccuracy = 0.7604
estimatedError = predictObjective(Results,x); estimatedAccuracy = 1 - estimatedError
estimatedAccuracy = 0.7580
テスト セットに対するモデルのパフォーマンスを評価します。結果から混同行列を作成し、混同行列におけるクラスの順序を指定します。
testAccuracy = 1 - loss(Mdl,creditTest,"Rating")testAccuracy = 0.7471
cm = confusionchart(creditTest.Rating,predict(Mdl,creditTest)); sortClasses(cm,["AAA","AA","A","BBB","BB","B","CCC"])
入力引数
名前と値の引数
出力引数
詳細
ヒント
データ セットのサイズ、指定した学習器の数、および選択した最適化手法によっては、
fitcautoの処理に時間がかかる場合があります。Parallel Computing Toolbox のライセンスがある場合、最適化を並列実行して計算を高速化することができます。これを行うには、
"HyperparameterOptimizationOptions",struct("UseParallel",true)を指定します。この構造体に追加フィールドを含めて、その他の最適化の側面を制御できます。HyperparameterOptimizationOptionsを参照してください。学習セットの観測値の数が原因 (たとえば 10,000 を超える場合) でベイズ最適化による
fitcautoの実行に長い時間がかかる場合は、代わりに ASHA 最適化によるfitcautoを使用することを検討してください。データ セットの観測値が多いと、多くの場合、ベイズ最適化よりも ASHA 最適化の方が優れた解を速く見つけます。ASHA 最適化を使用するには、"HyperparameterOptimizationOptions",struct("Optimizer","asha")を指定します。この構造体に追加フィールドを含めて、その他の最適化の側面を制御できます。特に、時間の制約がある場合は、HyperparameterOptimizationOptions構造体のMaxTimeフィールドを指定して、fitcautoを実行する秒数を制限します。
アルゴリズム
代替機能
使用するデータ セットに最適なモデルがわからない場合は、分類学習器アプリを代わりに使用することができます。このアプリを使用すると、さまざまなモデルについてハイパーパラメーターを調整して、パフォーマンスが最も高い最適化済みモデルを選択できます。分類学習器では、モデルのハイパーパラメーターを調整する前に特定のモデルを選択しなければなりませんが、最適化可能なハイパーパラメーターの選択とハイパーパラメーター値の設定をより柔軟に行うことができます。ただし、このアプリでは、最適化を並列実行することや、観測値の重みを指定すること、事前確率を指定すること、ASHA 最適化を使用することはできません。詳細は、分類学習器アプリのハイパーパラメーターの最適化を参照してください。
使用するデータに適合するモデルがわかっている場合は、対応するモデル近似関数を使用し、名前と値の引数
OptimizeHyperparametersを指定して、ハイパーパラメーターを調整することもできます。複数のモデルの結果を比較して最適な分類器を選択できます。このプロセスの例については、ベイズ最適化を使用したモデル選択の自動化への移行を参照してください。
参照
[1] Li, Liam, Kevin Jamieson, Afshin Rostamizadeh, Ekaterina Gonina, Moritz Hardt, Benjamin Recht, and Ameet Talwalkar. “A System for Massively Parallel Hyperparameter Tuning.” ArXiv:1810.05934v5 [Cs], March 16, 2020. https://arxiv.org/abs/1810.05934v5.
