Machine learning lecture2 note

Lecture 2

  linear regression,gradient descent and the normal equations(正规方程组？）

regression problem:continuous value

m: the number of raining example

X: input variables(features)

Y: target variable

(X,Y) is called one training example

for linear regression, h_{\theta}(x)=\theta_0+\theta_1 x_1+\theta_2 x_2+\cdots

"h" is called parameter, aim: use the training set to get the proper parameter

minimize {1\over2}\sum_{i=1}^n {(h_{\theta}(x^i)-y^i)^2}

one problem of gradient descent is that, a slightly alternate of the starting point may make you end up with a totally different local optimization.