1  定义

机器学习 (Machine Learning)：improving some performance measure with experience computed from data

2  应用举例

ML：an alternative route to build complicated systems

2.1  股票预测

 

2.2  图像识别

2.3  衣食住行

  

2.4  关键要素

在决定某些应用场景，是否适合使用机器学习时，常考虑以下三个要素：

1) exists some 'underlying pattern' to be learned, so 'performance measure' can be improved

2) but no programmable (easy) definition, so ML is needed

3) somehow there is data about the pattern, so ML has some 'inputs' to learn from

3  机器学习

下面以银行信用卡的申请为例，详细介绍机器学习的模型

3.1  申请者信息

  unknown pattern to be learned : "approve credit card good for bank?"

 

3.2  基本符号

  

3.3  机器学习过程

 

1)  目标函数

 

2)  学习模型

 

3)  学习过程

  ML： use data to compute hypothesis g that approximate target f

4  相关领域

4.1  与数据挖掘

  difficult to distuguish ML and DM in reality

Data Mining：use (huge) data to find property that is interesting

4.2  与人工智能

  ML is one possible route to realize AI

Artificical Intelligence：compute something that shows intelligent behavior

4.3  与统计学

  statistic has many useful tools for ML

  Statistics：use data to make interface about an unknown process

笔记资料

<机器学习基石> 林轩田，Lecture 1