Location Order Recovery in Trails with Low Temporal Resolution

TL;DR

This paper addresses the challenge of determining the correct sequence of object movements in trails with low temporal resolution, introducing a framework that effectively recovers location order despite broken points.

Contribution

It presents a novel three-phase framework using Markov transition networks to recover location order in trails with inadequate temporal resolution, handling broken points.

Findings

Framework efficiently recovers location order.

Correcting order significantly alters location criticality.

Method outperforms baseline approaches.

Abstract

Researchers who study object movement problems related to topics like traffic flow analysis, patient monitoring, and software operation, need to know the correct order in which objects move. Here, we use the term trail to refer to a series of movements by an object. This paper introduces a new missing data problem that occurs when analyzing trails where there is inadequate temporal resolution on the events. The temporal resolution is inadequate when an object, which can only be in one place at one time, appears in the data to be in two or more locations at once. We refer to this lack of resolution as a broken point. Broken points prevent us from knowing the correct order of movement. We propose a three-phase framework for recovering the location order. Based on the Markov transition network, we are able to find the route with the highest probability. Our results show that this framework can efficiently find the correct location order in trails with low temporal resolution. We also demonstrate that by correcting the location order, the criticality of locations can change significantly.