Linear Regression
regression and classification
| x \ y | continuous | discrete |
|---|---|---|
| continuous | regression | classification |
| discrete | regression | ？？？？ |
Simple linear regression
Least-squared fitting
Minimize the least square error
这里的 residual，就是对于单个点和计算得到的线性回归函数的误差，也就是相同x值下的两个y的差值。
Point estimation of β0 and β1
置信区间，估计不考，省略
Multiple linear regression
Least-squared fitting
some other calculation and parameters are in the ppt 4&5 p14