Online Program Home
My Program

Abstract Details

Activity Number: 635 - Advances in Machine Learning
Type: Contributed
Date/Time: Thursday, August 2, 2018 : 8:30 AM to 10:20 AM
Sponsor: Section on Statistical Learning and Data Science
Abstract #330663
Title: An Approximation to the Information Matrix of Hidden Markov Model
Author(s): Qing Ji* and Andrew Raim and Nagaraj Neerchal
Companies: University of Maryland, Baltimore County and U.S. Census Bureau and University of Maryland, Baltimore County
Keywords: Hidden Markov Model; Information Matrix

A hidden Markov model depends on an unobserved Markov process which determines the distribution of the observed outcome over time. The Fisher information matrix is often used to compute standard errors in likelihood-based inference, but a closed form of the matrix is not available for hidden Markov models. In this work, we propose an approximation to the FIM based on the join probability density function of response variable and the unobserved Markov process, and show it converges to the true information matrix under certain conditions.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program