Estimating an Activity Driven Hidden Markov Model

TL;DR

This paper introduces an activity-driven Hidden Markov Model to better infer human mobility patterns from mobile phone data, accounting for time-dependent activity levels influencing state transitions and emissions.

Contribution

It proposes a novel HMM framework incorporating time-dependent activity levels, enabling more accurate modeling of human mobility on sub-daily time scales.

Findings

Effective estimation of model parameters demonstrated

Improved inference of human mobility patterns

Applicable to mobile phone record analysis

Abstract

We define a Hidden Markov Model (HMM) in which each hidden state has time-dependent $\textit{activity levels}$ that drive transitions and emissions, and show how to estimate its parameters. Our construction is motivated by the problem of inferring human mobility on sub-daily time scales from, for example, mobile phone records.