BeSTAD: Behavior-Aware Spatio-Temporal Anomaly Detection for Human Mobility Data

2025-10-13

AI Summary

Traditional mobile anomaly detection primarily focuses on trajectory-level statistical outliers, failing to capture fine-grained deviations of individuals from their own historical behavioral patterns. To address this, we propose a behavior-aware spatiotemporal anomaly detection framework for large-scale, unlabeled human mobility data. First, we jointly model location semantics and periodic temporal patterns to learn interpretable, individual-specific mobility representations. Second, we introduce a behavior-cluster-aware mechanism to construct personalized behavioral profiles, enabling anomaly localization via cross-period behavioral comparison and semantic alignment. The method operates in a fully unsupervised manner, accurately identifying deviations from individual daily routines without labeled anomalies. It significantly enhances personalization and interpretability of detection results. Extensive experiments on real-world urban mobility datasets demonstrate its effectiveness and robustness across diverse scenarios.

Technology Category

• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search

Abstract

Traditional anomaly detection in human mobility has primarily focused on trajectory-level analysis, identifying statistical outliers or spatiotemporal inconsistencies across aggregated movement traces. However, detecting individual-level anomalies, i.e., unusual deviations in a person's mobility behavior relative to their own historical patterns, within datasets encompassing large populations remains a significant challenge. In this paper, we present BeSTAD (Behavior-aware Spatio-Temporal Anomaly Detection for Human Mobility Data), an unsupervised framework that captures individualized behavioral signatures across large populations and uncovers fine-grained anomalies by jointly modeling spatial context and temporal dynamics. BeSTAD learns semantically enriched mobility representations that integrate location meaning and temporal patterns, enabling the detection of subtle deviations in individual movement behavior. BeSTAD further employs a behavior-cluster-aware modeling mechanism that builds personalized behavioral profiles from normal activity and identifies anomalies through cross-period behavioral comparison with consistent semantic alignment. Building on prior work in mobility behavior clustering, this approach enables not only the detection of behavioral shifts and deviations from established routines but also the identification of individuals exhibiting such changes within large-scale mobility datasets. By learning individual behaviors directly from unlabeled data, BeSTAD advances anomaly detection toward personalized and interpretable mobility analysis.

Problem

• Detects individual-level anomalies in human mobility patterns
• Identifies behavioral deviations from personal historical movement data
• Models spatial context and temporal dynamics for fine-grained anomaly detection

Innovation

• Unsupervised framework modeling spatial-temporal mobility patterns
• Behavior-cluster-aware modeling with personalized activity profiles
• Semantically enriched representations integrating location meaning dynamics