In this section, we will train the (multi-class) SVM classifier with the MNIST training dataset and then use it to predict the labels of the images from the MNIST test dataset.
SVM is a pretty complex binary classifier that uses quadratic programming to maximize the margin between the separating hyper-planes. The binary SVM classifier is extended to handle multi-class classification problems using the 1-vs-all or 1-vs-1 technique. We are going to use scikit-learn's implementation, SVC(), with polynomial kernel (of degree 2) to fit (train) the soft-margin (kernelized) SVM classifier with the training dataset and then predict the labels of the test images using the score() function.
The following code shows how to train, predict, ...
