《learning from data》读书笔记---第一章: The Learning Problem

1 The Learning Problem

learning from data instead of analytic solution

1.1 Problem Setup

电影推荐系统：需要给电影评分，以确定是否给用户推荐此电影

模型基本步骤：基于之前的用户评分：1.构建向量描述电影 2.构建向量描述用户 3.计算这两个向量的相似度，预测评分

 It starts with random factors, then tunes these factors to make them more and more aligned with how viewers have rated movies before, until they are ultimately able to predict how viewers rate movies in general.

 1.1.1 Components of Learning 

There is a target to be learned. It is unknown to us. We have a set of examples generated by the target. The learning algorithm uses these examples to look for a hypothesis that approximates the target.  

1.1.2 A Simple Learning Model 

perceptron learning algorithm （PLA感知机算法）

 b与threshold有关，对于二维模型来说，分界线为：w1x1 + w2x2 + b = 0

PLA算法：1.初始化模型，w0->零向量 2.选择一个被误分类的记录用于更新w(t),其中t为迭代次数，w(t + 1) = w(t) + y(t)x(t)，直到不存在误分类记录，停止迭代。更新的物理意义： moves in the direction of class