李宏毅《机器学习》笔记打卡第二天

Regression

intput ===>function===>output

Our goal is find a best function.

Example Application

x(x_{cp},x_{s},x_{hp},x_{w},x_{h})====>f(宝可梦)===>cp after evolution (y)

A set of function :y=b+w*x_{cp}

there will have many function:

 f_{1}:y=10.0+9.0*x_{cp}

 f_{1}:y=9.8+9.2*x_{cp}

......

Linearmodel:y=b+\sum w_{i}x_{i}(x_{i}:input (x),w_{i}:weight)

Train data: input(x) output(y)

 Loss function L:

input: a function

output: how bad it is

L(f)=L(w,b)=\sum_{n=1}^{10}(y^n-(b+w*c^n_{cp}))^2

 

 

 

李宏毅《机器学习》笔记打卡第二天_第1张图片

 

你可能感兴趣的:(机器学习实战,机器学习,人工智能)