國家理論科學研究中心學術演講
NCTS Student's Seminar on Probability

　

In this semester, our student seminar in probability theory is research-oriented. We will introduce the audience some interesting topics popular in the field recently. As usual, students are invited to hold lectures on them. Our first choice of topic is Monte Carlo Markov Chain (MCMC) algorithm. The title of the paper is “General state space Markov Chain and MCMC algorithms”. This paper surveys various results about Markov chain on general state spaces and begins with an introduction to MCMC logarithm.Then the following are discussed: Sufficient Conditions for Geometric and Uniform Ergodicity. Quantitative Bounds on the Rate of Convergence to Stationarity. Central Limit Theorems- Necessary and Sufficient Conditions, via Poisson Equation or Direct Regeneration Constructions. Optimal Scaling and Weak Convergence Results for
Metropolis-Hastings Algorithm.One can download this paper from the website: http://probability.ca/jeff/.

　

  Speaker  :      陳與庭 (交通大學財金所碩士生)

                     

  Time      :         March 2 (Friday)  (I) AM 9:00 - 10:15     (II) AM 10:30 - 11:45

   

  Topic     :        General state space Markov chain and MCMC algorithms (I) (II)

               

  Place     :  

Lecture Room B,

National Center for Theoretical Sciences,

4th Floor, The 3rd General Building

National Tsing Hua University