ECS 256 Topic Outline and Reading List

This list may be overly optimistic. We will definitely get through HMMs and Queuing, hopefully more.

Notation:

• "PSB" means "Probability and statistics book," ECS 256 version -- BE SURE YOU ARE USING THIS VERSION OF THE BOOK, rather than, say the ECS 132 version or the ones for ECS 256 in previous years. Available here. (Note: published version lacks some chapters that we have here.)

• "SRC" means my book, Statistical Regression and Classificaton: from Linear Models to Machine Learning. This book is NOT required, but excerpts will be put online for you.
• "RecSys" means my recommender systems book (work in progress)

General preparation:

• Review of undergraduate probability, PSB Chapter 10.
• Linear algebra, PSB Appendix B.
• A cautionary tale, PSB Section 5.1.
• Conditional expectation as a random variable, SRC pp.52-54
• Quick introduction to the R language.

Discrete-Time Markov Chains

• Main material, PSB Chapters 6 and 12.
• Application: Google PageRank, RecSys, Chapter 12
• R packages: pageRank

Continuous-Time Markov Chains

• Preparation: The exponential distribution family, PSB Chapter 9.
• Main material, PSB Chapter 11.

Hidden Markov Chains

• Preparation:
• Statistical estimation, PSB, Chapter 22.
• Mixture distributions, PSB Chapter 24.
• Smoothing and the bias/variance tradeoff, PSB Chapter 25.

• Main material: TBA, mostly research papers.
• R packages: HMM
• Applications: NLP, time series, biosequences, etc.

Queuing Models

• Preparation: Transform methods, PSB Chapter 18
• Main material: PSB, Chapter 13
• Applications: Hospital management, communications networks, etc.
• R packages: queueing, simmer

Hazard Functions and Renewal Theory

• Main material: PSB, Chapters 14, 15
• R packages: chosen from CRAN Task View on Survival Analysis
• Application: Modeling Web acccess analysis, e.g. my paper
• Application: Modeling lost pay in job discrimination litigation, article by Pan and Gastwirth

Other Topics as Time Permits