フィルターのクリア

cost function in classification learner app

3 ビュー (過去 30 日間)
Artem Bykanov
Artem Bykanov 2018 年 5 月 7 日
編集済み: Stephan 2018 年 5 月 9 日
How is it possible to use a costfunction/ penalty function for a Support Vector Machine in the classification Learner App?
If it is not possible, is there a way to train a SVM manually for more than 2 classes?

採用された回答

Stephan
Stephan 2018 年 5 月 8 日
Hi,
a cost function is not available as far as i know. For more than two classes you could use the function
fitcecoc
see documentation here:
Best regards
Stephan
  2 件のコメント
Artem Bykanov
Artem Bykanov 2018 年 5 月 9 日
Thanks for your answer!
Now I have another problem: In the classificationLearner App I can export a trained Model. (It is a 1x1 struct)
How can generate the same trained model? Because if I try to classify with fitcecoc and validate with crossval I dont get this 1x1 struct.
Best regards
Stephan
Stephan 2018 年 5 月 9 日
編集済み: Stephan 2018 年 5 月 9 日
Hi,
when you choose the option "generate code" from the button export model in the classification learner app you get a good commented script with matlab-code doing exactly what you choosed in classification learner app. Then you can modify for your purposes as you want. If you do SVM model with more than 2 classes, i guess the function classification learner will use
fitcecoc
See here for further information:
Best regards
Stephan

サインインしてコメントする。

その他の回答 (0 件)

カテゴリ

Help Center および File ExchangeClassification Learner App 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