/Machine-Learning-Coursera

Programming assignments for ML

Primary LanguageMatlab

Machine-Learning-Coursera

COURSE WEBSITE：Machine Learning by Andrew Ng

Ex1. Linear Regression

PART TITLE SCORE
1 Warm up exercise 10 / 10
2 Compute cost for one variable 40 / 40
3 Gradient descent for one variable 50 / 50
4 Feature normalization 0 / 0
5 Compute cost for multiple variables 0 / 0
6 Gradient descent for multiple variables 0 / 0
7 Normal equations 0 / 0

Ex2. Logistic Regression

PART TITLE SCORE
1 Sigmoid function 5 / 5
2 Compute cost for logistic regression 30 / 30
3 Gradient for logistic regression 30 / 30
4 Predict function 5 / 5
5 Compute cost for regularized LR 15 / 15
6 Gradient for regularized LR 15 / 15

Ex3. Multi-class Classification and Neural Networks

PART TITLE SCORE
1 Regularized logistic regression 30 / 30
2 One-vs-all classifier training 20 / 20
3 One-vs-all classifier prediction 20 / 20
4 Neural network prediction function 30 / 30

Ex4. Neural Network Learning

PART TITLE SCORE
1 Feedforward and cost function 30 / 30
2 Regularized cost function 15 / 15
3 Sigmoid gradient 5 / 5
4 Neural net gradient function (backpropagation) 40 / 40
5 Regularized gradient 10 / 10

Ex5. Regularized Linear Regression and Bias/Variance

PART TITLE SCORE
1 Regularized linear regression cost function 25 / 25
2 Regularized linear regression gradient 25 / 25
3 Learning curve 20 / 20
4 Polynomial feature mapping 10 / 10
5 Cross validation curve 20 / 20

Ex6. Support Vector Machines

PART TITLE SCORE
1 Gaussian kernel 25 / 25
2 Parameters (C, sigma) for dataset 3 25 / 25
3 Email preprocessing 25 / 25
4 Email feature extraction 25 / 25

Ex7. K-Means Clustering and PCA

PART TITLE SCORE
1 Find closest centroids 30 / 30
2 Compute centroid means 30 / 30
3 PCA 20 / 20
4 Project data 10 / 10
5 Recover data 10 / 10

Ex8. Anomaly Detection and Recommender Systems

PART TITLE SCORE
1 Estimate gaussian parameters 15 / 15
2 Select threshold 15 / 15
3 Collaborative filtering cost 20 / 20
4 Collaborative filtering gradient 30 / 30
5 Regularized cost 10 / 10
6 Gradient with regularization 10 / 10