首页 > 代码库 > 机器学习笔记(Washington University)- Regression Specialization-week two

机器学习笔记(Washington University)- Regression Specialization-week two

1. Modeling seasonality

技术分享

w1 models the linear trend of the overall process.

w2 models the seasonal component sinusoid with a period of 12

and you do not know when the trend occurs so there exists a phase shift.

 

2. Regression model

技术分享

Now we are looking at the coeffiencts,for instanc,e the number of bedrroms, And 

what‘s the predicted change in the value of my house with everything else fixed.

Note:

the interpretation is in the context of what you have in the model. If the

square footage is fixed, the increse of the number of bedroom does not always

increase the value of the house.

 

3. RSS in matrix form and the gradient

技术分享

and the gradient of RSS is equal to:

技术分享

4.The close-form solution

 we can set the gradient to be zero to find the minimum:

技术分享

and we can get the W that we want.

 

5. Gradien descent

技术分享

 

机器学习笔记(Washington University)- Regression Specialization-week two