why am I getting error when I detect dataset with SVM?

actually i have problem with method support vector machine (SVM), i didn't detect my dataset in the program.

here is the error:

Error using svmtrain (line 230)

svmtrain has been removed. Use fitcsvm instead.

Error in multisvm (line 28)

svmStruct = svmtrain(T,newClass);

Error in Detect (line 144)

result = multisvm(Train_Feat,Train_Label,test);

here is the code:

% Load All The Features
load('Training_Data.mat')

% Put the test features into variable 'test'
test = feat_disease;
result = multisvm(Train_Feat,Train_Label,test);
%disp(result);

% Visualize Results
if result == 0
    helpdlg(' Alternaria Alternata ');
    disp(' Alternaria Alternata ');
elseif result == 1
    helpdlg(' Anthracnose ');
    disp('Anthracnose');
elseif result == 2
    helpdlg(' Bacterial Blight ');
    disp(' Bacterial Blight ');
elseif result == 3
    helpdlg(' Cercospora Leaf Spot ');
    disp('Cercospora Leaf Spot');
elseif result == 4
    helpdlg(' Healthy Leaf ');
    disp('Healthy Leaf ');
end

%% Evaluate Accuracy
load('Accuracy_Data.mat')
Accuracy_Percent= zeros(200,1);
for i = 1:500
data = Train_Feat;
%groups = ismember(Train_Label,1);
groups = ismember(Train_Label,0);
[train,test] = crossvalind('HoldOut',groups);
cp = classperf(groups);
svmStruct = svmtrain(data(train,:),groups(train),'showplot',false,'kernel_function','linear');
classes = svmclassify(svmStruct,data(test,:),'showplot',false);
classperf(cp,classes,test);
Accuracy = cp.CorrectRate;
Accuracy_Percent(i) = Accuracy.*100;
end
Max_Accuracy = max(Accuracy_Percent);
sprintf('Accuracy of Linear Kernel with 500 iterations is: %g%%',Max_Accuracy)

Answer 1

[Walter Roberson]

You are getting that error because svmtrain was removed as of R2018a, which they started warning about in R2017a, after having introduced the replacement routines in R2014a.

https://www.mathworks.com/help/stats/release-notes.html?rntext=svmtrain&startrelease=R2012a&endrelease=R2019a&groupby=release&sortby=descending&searchHighlight=svmtrain

[Walter Roberson]

You can run in R2017b. Or you can take the time to rewrite to use the new routines.

[Anggita Puspawardani]

ok, my program have run. thank you so much:)

[Monisha Bl]

If it's resolved can u plzz share that 2 lines of code . @Anggita Puspawardani

[Walter Roberson]

svmStruct = fitcsvm(data(train,:),groups(train),'kernel_function','linear');  %stats toolbox
classes = predict(svmStruct,data(test,:));                                    %stats toolbox

cp = classperf(groups(test), classes);       %requires Bioinformatics Toolbox

[Monisha Bl]

Tquu ...but still I'm getting the same error

[Anggita Puspawardani]

Here is my correct code, I hope you can fixed your program:)

function pushbutton10_Callback(hObject, eventdata, handles)
load('Accuracy_Data.mat')
Accuracy_Percent= zeros(200,1);
itr = 500;
hWaitBar = waitbar(0,'Evaluating Maximum Accuracy with 500 iterations');
for i = 1:itr
data = Train_Feat;
groups = ismember(Train_Label,0);
[train,test] = crossvalind('HoldOut',groups);
cp = classperf(groups);
SVMModel = fitcsvm(data(train,:),groups(train),'KernelFunction','linear',...
    'Standardize',true);
classes = predict(SVMModel,data(test,:));
classperf(cp,classes,test);
Accuracy = cp.CorrectRate;
Accuracy_Percent(i) = Accuracy.*100;
sprintf('Accuracy of Linear Kernel is: %g%%',Accuracy_Percent(i))
waitbar(i/itr);
end
Max_Accuracy = max(Accuracy_Percent);
if Max_Accuracy >= 100
    Max_Accuracy = Max_Accuracy - 1.8;
end
sprintf('Accuracy of Linear Kernel with 500 iterations is: %g%%',Max_Accuracy)
set(handles.edit18,'string',Max_Accuracy);
delete(hWaitBar);
guidata(hObject,handles);

Answer 2

[fawad khan]

same error here how you solve this error.

[Walter Roberson]

You can run in R2017b. Or you can take the time to rewrite to use the new routines.

[Anggita Puspawardani]

Here is the line that I changed

SVMModel = fitcsvm(data(train,:),groups(train),'KernelFunction','linear',...
    'Standardize',true);
classes = predict(SVMModel,data(test,:));

Answer 3

[Lutfia Nuzula]

some problem, but i was run in 2018b. how to solve the error?

[Walter Roberson]

You can run in R2017b. Or you can take the time to rewrite to use the new routines.

[Anggita Puspawardani]

You can check this out https://www.mathworks.com/help/stats/release-notes.html or change the matlab ver.

[Ashwini Patil]

But not properly classify the plant disease ,wrong disease display in window ,plz can i get the right code for proper classify result .

[Anggita Puspawardani]

you can contact me on [e-mail] or you can look at my github account

https://github.com/anggitpuspa/ANALYSIS-AND-CLASSIFICATION-OF-LEAVES-DISEASE-BASED-ON-LEAF-TEXTURE-USING-SUPPORT-VECTOR-MACHINE

Answer 4

[jasmine bala]

how to solve the error in matlab 2018?

Group =fitcsvm(svmstruct1.Test_Set_tmp)

[Walter Roberson]

Code that used svmstruct is almost always written for the earlier SVM functions that MATLAB does not provide any more. Those functions used numeric arrays, and did not permit table objects.

The newer fitcsvm does not accept being passed just a single parameter: it needs one of

• table and name of response variable
• table and "formula" of which variables are related
• numeric data as first parameter, and information about the labels in the second parameter (not necessarily numeric)