There are my codes of the exercises in Andrew Ng's Machine Learning course on Coursera. This project is to learn how to use Github and record my learning of Machine Learning.

Linear regression of Single and Multiple feature using Gradient Decent

Score: 100%

Logistic regression and Regularization using advanced optimization algorithm

Score: 100%

Logistic regression and Neural Networks for multi-classification (parameters of NN are provided)

Score: 80%
Error: Logistic regression is not suitable for dataset of any size

Score: 100%

Feedforward and Backpropgation of Neural Networks (with and without regularization)

Score: 75%
Error: When calculating regularization term of cost function, the parameters of biases are not ignored

Score: 100%

Use train set, cross-validation set and test set to choose lambda of Regularization, the features used in Polynomial regression and the size of train set

Score: 80%
Error: when using the sub-set of the train set, the size is not changed to the size of the sub-set

Score: 100%

Try linear kernel and Gaussian kernel of Support Vector Machine; try different C and sigma parameters of Gaussian kernel
Implement a simple spam classifier by using SVM

Score: 100%

Achieve K-means and Principle Component Analysis; apply K-means to image compression; apply PCA to face image compression

Score: 100%