I have an error with training SVM in Matlab. is there anyone can help me?

clear all
clc
ds=dataset('xlsfile','AD.xlsx');
feat1=ds.beta2L(1:100);
feat=ds.beta2T(1:100);
segmented_features = cell(1,3);
segmented_features =[feat1, feat];
 
[cA1,cH1,cV1,cD1] = dwt2(segmented_features,'db4');
[cA2,cH2,cV2,cD2] = dwt2(cA1,'db4');
[cA3,cH3,cV3,cD3] = dwt2(cA2,'db4');
 
DWT_feat = [cA3,cH3,cV3,cD3];
G = pca(DWT_feat);
whos DWT_feat
whos G
load trainset
xdata = meas;
 group = label;
 %svmStruct = svmtrain(xdata,group,'showplot',false);
 % species = svmclassify(svmStruct,feat)
 svmStruct1 = svmtrain(xdata,group,'kernel_function', 'linear');
 species = svmclassify(svmStruct1,feat,'showplot',false);
 classperf(cp,classes,test);
Accuracy_Classification = cp.CorrectRate.*100;
sprintf('Accuracy of Linear kernel is: %g%%',Accuracy_Classification)

**********************

Error using svmtrain (line 236)

Y must be a vector or a character array.

Error in newcodewithoutkmeans (line 22)

svmStruct1 = svmtrain(xdata,group,'kernel_function', 'linear');

[Walter Roberson]

what is class(group) and size(group)

Also which MATLAB version are you using?

[Vania krm]

Hello Walter,

class and size (group): 185*6 cell

matlab2017a

thanks

[Walter Roberson]

svmtrain() is only for binary training, where each sample is a member of one of two groups. You can use a cell array of character vectors, but the number of entries in that cell array needs to be the same as the number of rows in xdata and the cell needs to be a vector (1 x N or N x 1). Currently you have 186 x 6, which is not a vector (and which hints that your data might be multi-class)

[Vania krm]

Yes. Now my problem is solved. Very very thanks of your help. Just can I ask you that I use this code for accuracy based on Matlab codes of site but I have different accuracy in each time of run. How do I mention the final accuracy of system?

% Multiple runs for accuracy highest is 90%

groups = ismember(label,'AD ');

groups = ismember(label,'Nold');

[train,test] = crossvalind('HoldOut',groups);

cp = classperf(groups);

svmStruct = svmtrain(xdata(train,:),groups(train),'showplot',false,'kernel_function','linear');

classes = svmclassify(svmStruct,xdata(test,:),'showplot',false);

classperf(cp,classes,test);

Accuracy_Classification = cp.CorrectRate.*100;

sprintf('Accuracy of Linear kernel is: %g%%',Accuracy_Classification)

[Walter Roberson]

svmtrain uses random partitions . You would need to set the random seed to repeat the same results .