Cross-validated model parameters extraction in fitcecoc SVM linear kernel

Question

AMIR KHAN 2021-10-29

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1575088-cross-validated-model-parameters-extraction-in-fitcecoc-svm-linear-kernel

评论： AMIR KHAN 2021-10-30

Hi,

I am trying to extract the parameters from the cross-validated model of fitcecoc SVM with linear kernel. I want to extract the parameters which I could use for training and testing. I tried to look into the model but I could not find the bias/beta etc.

mdl = fitcecoc(Xs,Y,'Coding','onevsall','Learners',t);

CVMDL = crossval(mdl, 'kfold',5);

Error = kfoldLoss(CVMDL)

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Sulaymon Eshkabilov 2021-10-30

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1575088-cross-validated-model-parameters-extraction-in-fitcecoc-svm-linear-kernel#answer_820043

You should use these fcns instead of crossval:

kfoldPredict(), predict(), resubPredict()

See for more info: https://www.mathworks.com/help/stats/classreg.learning.partition.classificationpartitionedkernel.kfoldpredict.html

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

AMIR KHAN 2021-10-30

I want to use the parameters from the cross validated model. I want to fix these parameters to my model. Then I want to train my model on whole train set. Finally, test the trained model using test set.

请先登录，再进行评论。