Showing 1–2 of 2 results for author: Darwin, O

On the Sequential Probability Ratio Test in Hidden Markov Models

Authors: Oscar Darwin, Stefan Kiefer

Abstract: We consider the Sequential Probability Ratio Test applied to Hidden Markov Models. Given two Hidden Markov Models and a sequence of observations generated by one of them, the Sequential Probability Ratio Test attempts to decide which model produced the sequence. We show relationships between the execution time of such an algorithm and Lyapunov exponents of random matrix systems. Further, we give c… ▽ More We consider the Sequential Probability Ratio Test applied to Hidden Markov Models. Given two Hidden Markov Models and a sequence of observations generated by one of them, the Sequential Probability Ratio Test attempts to decide which model produced the sequence. We show relationships between the execution time of such an algorithm and Lyapunov exponents of random matrix systems. Further, we give complexity results about the execution time taken by the Sequential Probability Ratio Test. △ Less

Submitted 5 February, 2023; v1 submitted 28 July, 2022; originally announced July 2022.

Comments: 28 pages, 10 figures, submitted to CONCUR 2022

ACM Class: F.0; G.3

Equivalence of Hidden Markov Models with Continuous Observations

Authors: Oscar Darwin, Stefan Kiefer

Abstract: We consider Hidden Markov Models that emit sequences of observations that are drawn from continuous distributions. For example, such a model may emit a sequence of numbers, each of which is drawn from a uniform distribution, but the support of the uniform distribution depends on the state of the Hidden Markov Model. Such models generalise the more common version where each observation is drawn fro… ▽ More We consider Hidden Markov Models that emit sequences of observations that are drawn from continuous distributions. For example, such a model may emit a sequence of numbers, each of which is drawn from a uniform distribution, but the support of the uniform distribution depends on the state of the Hidden Markov Model. Such models generalise the more common version where each observation is drawn from a finite alphabet. We prove that one can determine in polynomial time whether two Hidden Markov Models with continuous observations are equivalent. △ Less

Submitted 27 September, 2020; originally announced September 2020.

Comments: 17 pages, 9 figures, Submitted to FSTTCS 2020

ACM Class: F.0; G.3