Hi Akash,

the error comes from these lines inside svmtrain.m:

% check group is a vector -- though char input is special...

if ~isvector(groupnames) && ~ischar(groupnames)

    error('Bioinfo:svmtrain:GroupNotVector',...

    'Group must be a vector.');

end

where groupnames variable is your input_data.X.

Notice that, the dimension of inputData should be at least 2, because you need to define at least 2 classes examples to train your classifier. I guess that your outputData vector will be the class labels,indicating whether they belong to class A or class B.

Can you check the sizes of both labels and input features vectors?

Hi Akash,

Jose is right. First thing to do is to check that your variable inout_data.X is a vector. It must be the same length as the number of rows in output_data.Y. May I clarify that output_data.Y is a matrix r x c, with r observations (rows) of c different variables - these variables will be the predictors used in your model.

i too had same problem earlier.... got suggestion to add path of libsvm in matlab....then executed....

had the same problem using ANN in matlab.  Try transposing your matrix.

hello dear

try declaring the class as a vector in the program.

All the best.

Dr.Indrajit Mandal

set path in the same folder where the libSVM-matlab at. it works! ^_^

plz show me how to add libSVM path

you can add the path in this way:- where your libsvm matlab have installed.

before adding the path please set the path :- home ->set path.

than add the path. Please follow one line code.

addpath('/home/gs/Desktop/AlexNet-Experiments-Keras-master/libsvm-3.23/matlab');