how to use multisvm

i want to use svm classifier for 4 class classification, and want to know how can i use svm for multiclass

Answer 1

[Purvaja]

Hi @shafna mtp,

I see that you wish to classify multiple classes using SVM. We can achieve this using “ClassificationECOC” function. You can perform the following steps:

• Assume some data with 4 classes for prediction
• Combine the classes into a single dataset

data = [class1; class2; class3; class4]; 

X = data(:, 1:2); % Features 

Y = data(:, 3); % Labels 

• Train the SVM model using fitcecoc

svmModel = fitcecoc(X, Y); 

• Display the trained model

disp(svmModel); 

• Make predictions on the training data

predictions = predict(svmModel, X); 

• Calculate accuracy

accuracy = sum(predictions == Y) / length(Y) * 100; 

fprintf('Accuracy: %.2f%%\n', accuracy); 

• Get predictions for some points

Predictions = predict(svmModel, Points); 

You can check another example by simply typing this command in your MATLAB command window:

openExample('stats/TrainAnECOCClassifierUsingSVMLearnersExample') 

For more clarification details on the functions used, you can refer “ClassificationECOC”, check out the following in documentation:

https://www.mathworks.com/help/releases/R2024a/stats/classificationecoc.html

https://www.mathworks.com/help/stats/support-vector-machine-classification.html

Hope this helps you!!