GTRSS: Graph-based Top-$k$ Representative Similar Subtrajectory Query

TL;DR

GTRSS introduces a graph-based framework for efficient top-k subtrajectory retrieval, significantly improving speed and accuracy over existing methods by using a dual-layer graph index and novel similarity metrics.

Contribution

This paper presents the first graph-based solution for top-k subtrajectory search, combining a dual-layer graph index with a new similarity metric and filtering techniques.

Findings

Achieves over 90% retrieval accuracy.

Up to two orders of magnitude speedup in query performance.

Significantly improves efficiency and accuracy over existing methods.

Abstract

Trajectory mining has attracted significant attention. This paper addresses the Top-k Representative Similar Subtrajectory Query (TRSSQ) problem, which aims to find the k most representative subtrajectories similar to a query. Existing methods rely on costly filtering-validation frameworks, resulting in slow response times. Addressing this, we propose GTRSS, a novel Graph-based Top-k Representative Similar Subtrajectory Query framework. During the offline phase, GTRSS builds a dual-layer graph index that clusters trajectories containing similar representative subtrajectories. In the online phase, it efficiently retrieves results by navigating the graph toward query-relevant clusters, bypassing full-dataset scanning and heavy computation. To support this, we introduce the Data Trajectory Similarity Metric (DTSM) to measure the most similar subtrajectory pair. We further combine R-tree and grid filtering with DTSM pruning rules to speed up index building. To the best of our knowledge, GTRSS is the first graph-based solution for top-k subtrajectory search. Experiments on real datasets demonstrate that GTRSS significantly enhances both efficiency and accuracy, achieving a retrieval accuracy of over 90 percent and up to two orders of magnitude speedup in query performance.