I am using the following script in MatlabR2015a:
CVO = cvpartition(y2,'k',10);
C=zeros(96,1)
for i = 1:CVO.NumTestSets
trIdx = CVO.training(i);
teIdx = CVO.test(i);
mdl = fitcnb(x1(trIdx, :),y2(trIdx, :),'Distribution', 'kernel')
[label, Posterior] = predict(mdl,x1(teIdx, :))
C(teIdx)=Posterior (:,2);
end
[X,Y,T,AUC] = perfcurve(y2,C,'true','NBoot', 1000); % Curva ROC para os dados de teste do crossval
c1=round(C);
CMat=confusionmat(y1,c1);
How can I save in my workspace the model from each iteration?
Many Thanks,
Elizabeth Vieira

 Réponse acceptée

Star Strider
Star Strider le 21 Jan 2016
You can save it as a cell array (note the curly brackets ‘{}’):
mdl{i} = fitcnb(x1(trIdx, :),y2(trIdx, :),'Distribution', 'kernel');
If you want to save the ‘mdl’ cell array after the loop, use the save function to save it to a .mat file. See the documentation for save for details.

3 commentaires

Stephen23
Stephen23 le 21 Jan 2016
Note that it is also recommended to preallocate cell arrays.
Elizabeth Vieira
Elizabeth Vieira le 22 Jan 2016
Many thanks!
Star Strider
Star Strider le 22 Jan 2016
My pleasure!

Connectez-vous pour commenter.

Plus de réponses (0)

Tags

Community Treasure Hunt

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

Start Hunting!

Translated by