[coursera machine learning] Week 1

1. machine learning 问题的分类：

Supervised Learning: right answers given in samples

　　　　　　　　　　　Regression: continuous result

　　　　　　　　　　　Classification: discrete valued output

Unsupervised Learning: learning about a dataset without correct answers

　　　　　　　　　　　　Clustering: divide dataset into groups

　　　　　　　　　　　　Non-clustering: separate different voices from a voice sample (cocktail party)

2. Model Representation:

training set -> learning algorithms -> hypothesis 

x -> hypothesis -> y

3. Cost Function:

m is the number of samples

4. Gradient Descent (not only for linear regression)

n is the number of features

minimization a function (ect. cost function)

the alpha is learning rate

all theta should be updated simultaneously.

5. Normal Equation Formula

comparison of gradient descent and normal equation formula.

normal equation is faster with less features.

gradient descent is faster with more features.