CAS MA 575: Linear Models (Fall 2013)
This is a post-introductory course to linear models, which focuses on
both principles and practice. We will cover simple and multiple linear regression,
weighted and generalized least squares, polynomials and factors,
transformations, regression diagnostics, variable selection, and a
selection from topics on extensions of linear models.
This course provides 4 credits, and lasts for one semester.
The prerequisites for this course are CAS MA 214, 242 (or 442), and
581; or consent of instructor.
Weisberg, S. (2005)
Applied Linear Regression, Third Edition.
John Wiley and Sons, New York.
This book contains a number of typos and errors. Please consult the
published by the author.
- Xinyu Kang:
- Mondays from 1pm to 2pm in office MCS 152.
- Tuesdays from 12:30pm to 1:30pm in office MCS 152.
- Cedric Ginestet:
- Wednesdays from 3pm to 5pm in office MCS 229.
- Week 1:
Introduction and Preliminaries
- Week 2:
Scatterplots, LOESS, OLS Estimation
- Week 3:
Properties of OLS Estimators, MSE, F-test
- Week 4:
Tests for Individual
Estimators, Multiple Regression
- Week 5:
Hat Matrix, ANOVA Table
- Week 6:
MLE, Gauss-Markov, Weighted
- Week 7:
Midterm Review Session
- Week 8:
Unbiasedness of OLS Variance Estimator
- Week 9:
Bootstrap, Permutation Tests,
- Week 10:
Delta Method, Mixed Effects
- Week 11:
Residual Diagnostics, Leverages, Data Transformation
- Week 12:
Outliers, Influence, Cook's Distance, QQ-plots
- Week 13:
Variable Selection, Model Comparison, AIC, BIC
- Week 14:
Regularization, Ridge regression, Lasso
- Week 15:
Final Review Session