Efficient inference of hidden Markov models from large observation sequences

Benjamin W. Priest; George Cybenko

doi:10.1117/12.2230587

12 May 2016 Efficient inference of hidden Markov models from large observation sequences

Benjamin W. Priest, George Cybenko

Proceedings Volume 9825, Sensors, and Command, Control, Communications, and Intelligence (C3I) Technologies for Homeland Security, Defense, and Law Enforcement Applications XV; 98250V (2016) https://doi.org/10.1117/12.2230587
Event: SPIE Defense + Security, 2016, Baltimore, MD, United States

Abstract

The hidden Markov model (HMM) is widely used to model time series data. However, the conventional Baum- Welch algorithm is known to perform poorly when applied to long observation sequences. The literature contains several alternatives that seek to improve the memory or time complexity of the algorithm. However, for an HMM with N states and an observation sequence of length T, these alternatives require at best O(N) space and O(N²T) time. Given the preponderance of applications that increasingly deal with massive amounts of data, an alternative whose time is O(T)+poly(N) is desired. Recent research presents an alternative to the Baum-Welch algorithm that relies on nonnegative matrix factorization. This document examines the space complexity of this alternative approach and proposes further optimizations using approaches adopted from the matrix sketching literature. The result is a streaming algorithm whose space complexity is constant and time complexity is linear with respect to the size of the observation sequence. The paper also presents a batch algorithm that allow for even further improved space complexity at the expense of an additional pass over the observation sequence.

Citation Download Citation

Benjamin W. Priest and George Cybenko "Efficient inference of hidden Markov models from large observation sequences", Proc. SPIE 9825, Sensors, and Command, Control, Communications, and Intelligence (C3I) Technologies for Homeland Security, Defense, and Law Enforcement Applications XV, 98250V (12 May 2016); https://doi.org/10.1117/12.2230587

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available