Dynamic modeling of trajectory patterns using data mining and reverse engineering

This paper claims that meaningful patterns can only be extracted from trajectories if the geographic space where trajectories are located is considered, and proposes a reverse engineering framework for mining and modeling semantic trajectory patterns.

Abstract

The constant increase of moving object data imposes the need for modeling, processing, and mining trajectories, in order to find and understand the patterns behind these data. Existing works have mainly focused on the geometric properties of trajectories, while the semantics and the background geographic information has rarely been addressed. We claim that meaningful patterns can only be extracted from trajectories if the geographic space where trajectories are located is considered. In this paper we propose a reverse engineering framework for mining and modeling semantic trajectory patterns. Since trajectory patterns are data dependent, they may not be modeled in conceptual geographic database schemas before they are known. Therefore, we apply data mining to extract general trajectory patterns, and through a new kind of relationships, we model these patterns in the geographic database schema. A case study shows the power of the framework for modeling semantic trajectory patterns in the geographic space.