Week 1 Course Introduction

1.      Machine learning definition:

“A computer programis said to learn from experience E with respect to some task T and someperformance measure P, if its performance on T, as measured by P,improves with experience E.”

 

2.      Machine learning algorithms:

1)      Supervised learning

a.      Regression: Predict continuous valued output (price)

b.      Classification: Discrete valuedoutput (0 or 1)

2)      Unsupervised learning

3)      Reinforcement learning, recommender systems

 

3.      Linear regression with one variable

1)     Hypothesis: h_{\Theta }(x)=\Theta _{0}+\Theta _{1}x

Parameters:

Cost function:  

Goal: minimize

2)     Gradient descent

Repeat until convergence {}

a.      Small , slowconvergence; Large , maybefail or diverge.

b.      Fixed , couldconverge to local minimums too. No need to decrease.

3)   Gradient descent for linearregression

Repeat until convergence

4)      “Batch” Gradient Descent

Each step. All the training examples are used.