Implement naive bayesian and a spam email classifier using naive bayesian.

Implement some linear models, including Linear regression, logistic regression, perceptron, SVM and their testing. Also, a regularization version of these linear models. I choose the lambda in regularization using corss-validation.

Implement a one hidden layer neural network and its feedforward and backpropagation process and a weight decay and momentum version of it.

Implement KNN and use KNN to recognize CAPTCHA of my school website.

Implement K-means and use K-means in picture compression

Implement spectral clustering and use it in news clustering.

Implement PCA and use it to rotate a roteted CAPTCHA picture to horizontal direction.
Use PCA and KNN to recognize faces.