This project consists of 4 parts :
- First part is implemetation of binary SVM for two class recognition. Code for this is found in Step0.m.:
- Now the other three parts are implementation of Multiclass SVM using One vs All extension(Step1.m),
One vs One extension(Step2.m),error correcting coding(Step3.m)
- Each of the above files will perform following functions:
- Load the corresponding dataset and invokes SVM.m which implements the Binary SVM classifier.
- Predict the classifier obtained on testing dataset and displays the accuracy on it.
- Above two steps are performed for different cost functions.