Logical Hidden Markov Models

TL;DR

Logical Hidden Markov Models extend traditional HMMs to handle structured logical atoms, enabling more expressive sequence modeling, with formal inference solutions and experimental validation in bioinformatics.

Contribution

Introduces LOHMMs, a novel extension of HMMs for structured data, along with algorithms for inference and parameter estimation.

Findings

Effective evaluation and inference algorithms for LOHMMs

Successful application to bioinformatics problems

Demonstrated improved modeling of structured sequences

Abstract

Logical hidden Markov models (LOHMMs) upgrade traditional hidden Markov models to deal with sequences of structured symbols in the form of logical atoms, rather than flat characters. This note formally introduces LOHMMs and presents solutions to the three central inference problems for LOHMMs: evaluation, most likely hidden state sequence and parameter estimation. The resulting representation and algorithms are experimentally evaluated on problems from the domain of bioinformatics.