How to merge two .mat files?

I am trying to do svm training https://in.mathworks.com/matlabcentral/fileexchange/55098-plant-leaf-disease-detection-and-classification-using-multiclass-svm-classifier and https://in.mathworks.com/matlabcentral/fileexchange/55107-brain-mri-tumor-detection-and-classification in the above two link when i load the training data-set i get two or more .mat file Example in the first link when i load Training_Data it open Train_feat and Train_label...how can i do this? but before this what am trying is... i have one folder who have 3 types of training image each type have 50 image, totally 150 training image(3 class) i add one row which count the no of row, class one will have record from row 1-50, second 51-100, third 101-150 while testing if k value falls from this it will display the class label of that training image, but the out put is not too good...so i want to try the above two link examples..thank you sorry and for this boring write.

Answer 1

[Walter Roberson]

s1 = load('FirstFile.mat');
s2 = load('SecondFile.mat');
fn1 = fieldnames(s1);
fn2 = fieldnames(s2);
if any(ismember(fn1, fn2))
   error('Those files have at least one variable with the same name, cannot merge them');
end
for K = 1 : length(fn2)
  thisfield = fn2{K};
  s1.(thisfield) = s2.(thisfield);
end

save('MergedFile.mat', '-struct', 's1');

The output MergedFile.mat will contain all of the variables that were in each of the other two files.

... But I have no idea why you think merging .mat files will be useful in your situation.

[Melaku Eneayehu]

thank you sir for your quick answer.am start training multi-class svm, i plan one mat file will contain the label, i try one method that is...i extract feature and save in ,mat file as i mentioned i have 3 classes, and 150 row(extracted feature images) after doing this when i test i try to search similarity on row level, is similarity is close to 1-50 row it falls in to 1st class, if it falls from 51-100 it will falls in to 2nd class...and so on..but the out put is not too good...so i try another training method. i hope you get me....in short i am try training multi-class(3 class) svm based on GLCM and other texture feature totally that have 14 features, after preparing data-set i got difficult in training phase.

[Walter Roberson]

Merging mat files is not going to somehow give you better fitting results than the vague "output is not good" that you mention.