8_ Polynomial regression and model generalization

8_ Polynomial regression and model generalization (Polynomial Regression and Model Generalization)

In this chapter , We're going to be dealing with nonlinear problems . We will study the idea of polynomial regression , Use linear regression to solve nonlinear problems .
further , We will extend perhaps the most important question in machine learning ： Model generalization problem .
We will further explore what is under fitting , What is over fitting , How to detect under fitting and over fitting . What is cross validation , What is model regularization . It sounds like a mouthful Ridge and Lasso What the hell are they …

8-1 What is polynomial regression

8-2 scikit-learn Polynomial regression in and Pipeline

8-3 Over fitting and under fitting

8-4 Why training data sets and test data sets

8-5 The learning curve

8-6 Validation datasets and cross validation

8-8 Model generalization and ridge regression

summary ： Model regularization greatly improves the generalization ability of the model , The principle behind it is Because the model coefficient of over fitting is too large , Add an additional item , Give Way MSE The coefficient of part can be smaller .

8-9 LASSO

8-10 L1, L2 And elastic networks