While working with the Gaussian classifier in problem set 3, we were unable to utilize the information given in the covariance matrix and assumed the covariance matrix to be identity. This time we will apply PCA on the training data to reduce the dimensionality of our feature space