Another look at Bayes map iterated filtering

作者：Nguyen Dao^*

来源：Statistics & Probability Letters, 2016, 118: 32-36.

DOI：10.1016/j.spl.2016.05.013

• Iterated filtering
• Bayes map
• Hidden Markov model
• Stochastic approximation
• Time series

A new proof of convergence of Bayes map iterated filtering algorithm, based on super martingale inequality, is proposed. Relying on log concavity assumption, this approach with verifiable conditions is simpler than the previous approach and is generalizable to more sophisticated algorithms. The assumption could be relaxed at the cost of more technical details.

• 出版日期2016-11