03_regression.md

Section 03: Regression

Regression is important because it is stable, interpretable and widely understood and accepted, and easy to deploy.

Class Materials

• Instructor notes

Linear Regression

• Overview of interpreting linear regression

• The Bias/Variance Tradeoff

• Advanced notes

• EM linear regression example

• H2o linear regression example

• Basic gradient descent example

• Kaggle Advanced Regression Contest starter kit

---

Logistic Regression and GLM

• Overview of logistic regression - Blackboard electronic reserves

• Overview of interpreting logistic regression

• Confusion Matrix

• Overview of Penalized GLM

• EM logistic regression example

• H2o logistic regression example

• Overview of assessing binary classifiers

  • Interpreting assessment measures
  • Assessment measure workbook

• Assignment

• Assignment Key

Sample Quizzes

• Sample quiz 3.1 - bias/variance tradeoff
• Sample quiz 3.2 - linear regression
• Sample quiz 3.3 - logistic regression and assessment

Quiz Keys

• Linear regression
• Logistic regression

Supplementary References

• Introduction to Statistical Learning
  Sections 2.2, 3.1 - 3.3, and 4.3

• Elements of Statistical Learning
  Chapters 3, 4 and 7

• Generalized Linear Modeling with H2o

• Logistic regression in Enterprise Miner - Blackboard electronic reserves

• UCLA IDRE: How do I interpret odds ratios in logistic regression?

• Elastic Net Regression
  By Hastie and Zhou, 2005