352 – Contributed Oral Poster Presentations: Section on Bayesian Statistics
Parallel Particle Learning for Bayesian Financial Data Analysis
Hiroaki Katsura
Keio University
Kenichiro McAlinn
Keio University
Teruo Nakatsuma
Keio University
Posterior simulation for Bayesian inference using particle �lters and particle learning algorithms have proven to be successful in various �elds, including �nance. However, these particle based methods for posterior simulation are, by nature, computationally strenuous and time consuming. With the recent development of fast and inexpensive devices for parallel computing, such as general purpose graphic processing units (GPGPU), in mind, we have developed a new algorithm for particle �lter that is fully parallelized.