STAT 396
Finite Markov Chains and Monte-Carlo Methods
3.0 credits
Course Description
Finite Markov chains; stationary distributions; time reversals; classification of states; classical Markov chains; convergence in total variation distance and L2; spectral analysis; relaxation time; Monte Carlo techniques: rejection sampling, Metropolis-Hastings, Gibbs sampler, Glauber dynamics, hill climb and simulated annealing; harmonic functions and martingales for Markov chains. Prerequisite: a minimum grade of 2.0 in MATH 208; a minimum grade of 2.0 in either MATH 394/STAT 394, CSE 312, or STAT 340; and a minimum grade of 2.0 in either STAT 395/MATH 395, STAT 341, or STAT 391. Offered: jointly with MATH 396; Sp.