HW2: Exercises: From Marin and Robert, Due: 15 June 2011, 10:00
2.14, 2.15, 2.19
Further programming exercises are here.
HW1: Exercises: From Marin and Robert, Due: 8 June 2011, 10:00
1.2, 2.4, 2.5, 2.6, 2.7, 2.8, 2.9, 2.10, 2.12, 2.13,
Deliver in class as a printout.
You may discuss solution startegies with other students but the solution you deliver in must be genuinely yours.
You must deliver at least 3 questions for each homework, late submissions get only 50%, after a week 0%.
Grading: Each question, regardless of the difficulty, will be worth 1 pt. Partial solutions get 0.5.
Format: The first page must contain the list of questions that you are delivering. Suppose you could solve only 1.2, 2.4 and 2.10. Then, your first page must contain a table containing the following information:
Your Name
HW number
Delivery date
Q# — – Grade
1.2
2.4
2.10
Total
This Course will be an overview of Modern Bayesian Statistical Methods. We assume only an elementary exposure to probability. As is, the topics should be suitable for CMPE, EE, IE, Maths, CHE graduate students and anyone who is interested in analysis of datasets using modern techniques.
We assume elementary programming skills with Matlab and/or R. The focus will be on data analysis, not on program or system design.
CMPE58R has some overlap with the topics covered in CMPE 58K (Bayesian Statistics and Machine Learning) and CMPE 58N, Monte Carlo methods for Scientific computing. However, unlike these courses, the focus is on basic statistical concepts and models. We will contrast Bayesian methods and classical frequentist statistics. We won't cover Bayesian networks or hidden Markov models and will use Monte Carlo methods mostly as a computational tool.
Jean-Michel Marin, Christian P. Robert,
Bayesian Core: A Practical Approach to Computational Bayesian Statistics
Springer 2010
Course material, Slides, codes and datasets
William Fitzgerald, Course Notes, Dept. of Eng., University of Cambridge
Classical estimation - Fisher Info and CR bounds with examples
Bayesian methods - marginal estimators etc, priors, max ent to assign probabilities, numerical methods, Gibbs sampling etc
sufficient statistics, Neyman-Fisher factorization theorem - lots of examples
Stochastic processes - Markov processes, Chapman-Kolmogorov eqn, Fokker-Planck eqn, Brownian motion
Discrete event systems - Poisson processes, queueing theory
A. Taylan Cemgil, Cmpe, Bogazici
Guest Lecturer, 13 July - 8 August
William Fitzgerald, Prof. Dr., Dept. of Engineering, University of Cambridge
Will be based on (tentatively) 4,5 Homework sets, including some theoretical and programming assignments.
No midterm, no final.