MTH511A - 2020
Statistical Simulation and Data Analysis
This is the course webpage for MTH 511: Statistical Simulation and Data Analysis from 2020 Semester-I.
The course will be run completely online through mooKIT and live tutorial sessions. Due to bandwidth issues, the videos will also be on Youtube.
Information:
- 5:00pm - 6:00pm M,Th (live)
- R resources:
- Download R from CRAN website
- Download RStudio. (Optional, feel free to use the standard RScript or Terminal).
- R help: Video, forum, best resource.
Course Outline
Wk | Lectures | Topics | Resources |
---|---|---|---|
1 | Lec 1 | Introduction to Monte Carlo | Notes, R code, Video |
Lec 2 | Pseudo-random number generator | Notes, R code, Video | |
2 | Lec 3 | Inverse transform method (discrete) | Notes, Video 1, Video 2 |
Lec 4 | Accept-rejection method (discrete) | Notes, Code, Video | |
Lec 5 | Composition methods and Bernoulli factories | Notes, Video | |
3 | Lec 6 | Inverse transform method (continuous) | Notes, Video |
Lec 7 | Accept-reject method (continuous) | Notes, Code, Video | |
Lec 8 | Box-Muller and Ratio-of-Uniforms method | Notes 1, Notes 2, Code 1, Code 2, Video | |
4 | Lec 9 | Accept-Reject: more examples | Notes, Video |
Lec 10 | Miscellaneous topics in sampling | Notes, Code, Video | |
Lec 11 | Simple Importance Sampling | Notes, Code, Video | |
5 | Lec 12 | Simple importance sampling | Notes, Code, Video |
Lec 13 | Weighted importance sampling | Notes, Video | |
6 | Lec 14 | MLE and (ridge) regression | Notes, Video |
Lec 15 | Newton Raphson’s method | Notes, Code, Video | |
Lec 16 | Gradient Ascent | Notes, Code, Video | |
7 | – | Midsem Exams | |
8 | Lec 17 | MM algorithm | Notes, Code, Video |
Lec 18 | EM algorithm and mixture of Gaussians | Notes, Code, Video | |
Lec 19 | EM and Monte Carlo EM | Notes, Code, Video | |
9 | Lec 20 | Loss functions | Notes, Code, Video |
Lec 21 | Cross-validation | Notes, Code, Video | |
Lec 22 | Bootstrapping | Notes, Code, Video | |
10 | Lec 23 | Stochastic Gradient Descent | Notes, Code, Video |
Lec 24 | Simulated Annealing | Notes, Code, Video | |
Lec 25 | A Recap of Data Analysis Methods | Notes, Video | |
11 | – | Mini-Project Week | |
12 | Lec 26 | Introduction to Bayesian Models | Notes, Code, Video |
Lec 27 | Accept-Reject for Bayes | Notes, Code, Video | |
Lec 28 | Bayesian Linear Regression | Notes, Code, Video | |
13 | Lec 29 | Introduction to MCMC | Notes, Code, Video |
Lec 30 | Metropolis-Hastings Examples | Notes, Code, Video | |
Lec 31 | Bayesian Logistic Regression | Notes, Code, Video |
References:
- Sampling from Distributions:
- “Simulation” by Sheldon M. Ross (Academic Press, Fourth Edition), 2006, Chaps. 1-5.
- “Non-Uniform Random Variable Generation” by Luc Devroye. Online book
- Bernoulli Factories: “Designing perfect simulation algorithms using local correctness” arxiv
- Ratio-of-Uniforms method Paper, Resource 1, Resource 2
- Maximum Likelihood Estimation
- “Statistical Inference” by Casella and Berger.
- MLE estimation notes at UChicago link
- Regression
- “Applied Linear Regression” by Sanford Weisberg
- Elements of Statistical Learning by Hastie, Tibshirani, and Friedman Link
- Optimization
- EM Algorithm
- EM Algorithm proof pdf
- EM Algorithm notes 1 Link
- EM Algorithm notes 2 Link
- EM and MCEM Pages 113 - 135
- Cross-validation
- Bootstrap
- Larry Wasserman Notes
- Book non-pdf link
- MCMC
- Linchpin Accept-Reject
- My notes here
- Importance Sampling