1. Machine Learning CUNY Graduate Center Lecture 6: Linear Regression II

2. Extension to polynomial regression 1

3. Polynomial regression is the same as linear regression in D dimensions 2

4. Generate new features 3 Standard Polynomial with coefficients, w Risk

5. Generate new features 4 Feature Trick: To fit a D dimensional polynomial, Create a D-element vector from x i Then standard linear regression in D dimensions

6. How is this still linear regression? The regression is linear in the parameters, despite projecting x i from one dimension to D dimensions. Now we fit a plane (or hyperplane) to a representation of x i in a higher dimensional feature space. This generalizes to any set of functions 5

7. Basis functions as feature extraction These functions are called basis functions. –They define the bases of the feature space Allows linear decomposition of any type of function to data points Common Choices: –Polynomial –Gaussian –Sigmoids –Wave functions (sine, etc.) 6

8. Training data vs. Testing Data Evaluating the performance of a classifier on training data is meaningless. With enough parameters, a model can simply memorize (encode) every training point To evaluate performance, data is divided into training and testing (or evaluation) data. –Training data is used to learn model parameters –Testing data is used to evaluate performance 7

9. Overfitting 8

10. 9

11. Overfitting performance 10

12. Definition of overfitting When the model describes the noise, rather than the signal. How can you tell the difference between overfitting, and a bad model? 11

13. Possible detection of overfitting Stability –An appropriately fit model is stable under different samples of the training data –An overfit model generates inconsistent performance Performance –A good model has low test error –A bad model has high test error 12

14. What is the optimal model size? The best model size generalizes to unseen data the best. Approximate this by testing error. One way to optimize parameters is to minimize testing error. –This operation uses testing data as tuning or development data Sacrifices training data in favor of parameter optimization Can we do this without explicit evaluation data? 13

15. Context for linear regression Simple approach Efficient learning Extensible Regularization provides robust models 14

16. Linear Regression Identify the best parameters, w, for a regression function 15

17. Overfitting Recall: overfitting happens when a model is capturing idiosyncrasies of the data rather than generalities. –Often caused by too many parameters relative to the amount of training data. –E.g. an order-N polynomial can intersect any N+1 data points 16

18. Dealing with Overfitting Use more data Use a tuning set Regularization Be a Bayesian 17

19. Regularization In a linear regression model overfitting is characterized by large weights. 18

20. Penalize large weights Introduce a penalty term in the loss function. 19 Regularized Regression (L2-Regularization or Ridge Regression)

21. Regularization Derivation 20

22. 21

23. Regularization in Practice 22

24. Regularization Results 23

25. More regularization The penalty term defines the styles of regularization L2-Regularization L1-Regularization L0-Regularization –L0-norm is the optimal subset of features 24

26. Curse of dimensionality Increasing dimensionality of features increases the data requirements exponentially. For example, if a single feature can be accurately approximated with 100 data points, to optimize the joint over two features requires 100*100 data points. Models should be small relative to the amount of available data Dimensionality reduction techniques – feature selection – can help. –L0-regularization is explicit feature selection –L1- and L2-regularizations approximate feature selection. 25

27. Bayesians v. Frequentists What is a probability? Frequentists –A probability is the likelihood that an event will happen –It is approximated by the ratio of the number of observed events to the number of total events –Assessment is vital to selecting a model –Point estimates are absolutely fine Bayesians –A probability is a degree of believability of a proposition. –Bayesians require that probabilities be prior beliefs conditioned on data. –The Bayesian approach “is optimal”, given a good model, a good prior and a good loss function. Don’t worry so much about assessment. –If you are ever making a point estimate, you’ve made a mistake. The only valid probabilities are posteriors based on evidence given some prior 26

28. Bayesian Linear Regression The previous MLE derivation of linear regression uses point estimates for the weight vector, w. Bayesians say, “hold it right there”. –Use a prior distribution over w to estimate parameters Alpha is a hyperparameter over w, where alpha is the precision or inverse variance of the distribution. Now optimize: 27

29. Optimize the Bayesian posterior 28 As usual it’s easier to optimize after a log transform.

30. Optimize the Bayesian posterior 29 As usual it’s easier to optimize after a log transform.

31. Optimize the Bayesian posterior 30 Ignoring terms that do not depend on w IDENTICAL formulation as L2-regularization

32. Context Overfitting is bad. Bayesians vs. Frequentists –Is one better? –Machine Learning uses techniques from both camps. 31

33. Next Time Logistic Regression 32