Finding p-value of coefficients of a trained model

5 ビュー (過去 30 日間)
Muhammad Fazalul Rahman
Muhammad Fazalul Rahman 2018 年 4 月 24 日
I've trained a model using the "Regression Learner" app (GPR with Matern5/2 kernel), and I exported the model using the "Export Model" option. How do I obtain the coefficients and their corresponding p-values from the model? To re-learn the model it would take about 24 hours.

回答 (1 件)

Ahmet Cecen
Ahmet Cecen 2018 年 4 月 24 日
Just load the model then explore using dot notation:
M.RegressionGP
RegressionGP
PredictorNames: {'column_1'}
ResponseName: 'Y'
CategoricalPredictors: []
ResponseTransform: 'none'
NumObservations: 10
KernelFunction: 'Matern52'
KernelInformation: [1×1 struct]
BasisFunction: 'Constant'
Beta: 0.4994
Sigma: 0.2928
PredictorLocation: 0.4946
PredictorScale: 0.2321
Alpha: [10×1 double]
ActiveSetVectors: [10×1 double]
PredictMethod: 'Exact'
ActiveSetSize: 10
FitMethod: 'Exact'
ActiveSetMethod: 'Random'
IsActiveSetVector: [10×1 logical]
LogLikelihood: -1.9069
ActiveSetHistory: []
BCDInformation: []
  2 件のコメント
Ahmet Cecen
Ahmet Cecen 2018 年 4 月 24 日

Also here is an explanation of what each field is: Reference

Muhammad Fazalul Rahman
Muhammad Fazalul Rahman 2018 年 4 月 24 日
But is there a way to get the p-values of the regression coefficients?

サインインしてコメントする。

カテゴリ

Help Center および File ExchangeGaussian Process Regression についてさらに検索

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by