Regression
step1:model
y=b+w*x cp
step2:goodness of function
loss function:how bad the function is
![](https://img.haomeiwen.com/i6196923/db93473fffc9f84e.png)
step3:best function
令Lf最小
![](https://img.haomeiwen.com/i6196923/480c54932222b2db.png)
GRADIENT DESCENT:协助计算偏微分
![](https://img.haomeiwen.com/i6196923/e3f916dad6658548.png)
![](https://img.haomeiwen.com/i6196923/0d9ea17ebdbbc7d1.png)
Overfitting:A more complex model does not always lead to better performance on testing data.
hidden factor->
back to step1:redesign the model
back to step2:regularizaton
越平滑,对noise越不敏感。training data误差越小,但testing data误差不一定
![](https://img.haomeiwen.com/i6196923/e9962c7ebca67246.png)
网友评论