Introduction - If you have any usage issues, please Google them yourself
1, the maximum likelihood estimate of 0 to 24 try-order polynomial fitting, with 5-fold cross-validation to select the best model (polynomial order and coefficients are given in Figure similar courseware), and draw the best model Fitting renderings (similar to Figure 1, a blue dot to the training sample, the test sample is red dot, the green line is the model prediction), the test error of the model are given. 2, ridge regression polynomial order of 24, the regular coefficient ranges λ is exp (-19) to exp (20), and with the use of 5-fold cross-validation to select the best model. The experimental results with a requirement.