CS 542 Machine Learning course project This project is divided into 2 parts
- Train and test an SVM classifier with non-linear kernel, then use the classifier to compare multi-class classification performance of two different voting schemes:
- One vs. the rest
- One vs. one
- Use the same “one vs. one” classifiers in a DAGSVM approach
The classifiers are trained and tested by images of handwritten digits data set that represents a subsampling of the full MNIST data set available in http://yann.lecun.com/exdb/mnist/