Beyond Long Tail POIs: Transition-Centered Generalization for Human Mobility Prediction

2026-05-07

AI Summary

Human mobility prediction suffers from a long-tailed distribution at the transition level: even for popular points of interest (POIs), observed origin–destination transition records remain highly sparse, hindering model generalization. This work identifies this bottleneck as combinatorial sparsity in transitions and proposes RECAP, a novel framework that reconstructs sparse transitions by leveraging a global multi-hop transition graph and revisitation evidence extracted from user trajectories. To mitigate overfitting and enhance compositional generalization, RECAP incorporates a hot-transition-preserving training strategy. Evaluated on multiple real-world datasets, RECAP significantly improves next-POI prediction performance, with particularly pronounced gains on tail transitions.

Abstract

Human mobility prediction forecasts a user's next Point of Interest (POI) from historical trajectories, supporting applications from recommendation to urban planning. Recent studies have recognized the problem with long-tail POIs in human mobility prediction, which are POIs with few visit records, making new visits to such POIs difficult to predict. Our analysis shows that many predictions fail even for visits to popular POIs. The underlying cause is often transition-level sparsity: the corresponding source-destination transition appears rarely, or never appears, in the training set. We therefore argue that a core bottleneck in human mobility prediction lies in transition-level long-tail generalization. We formulate this problem as compositional generalization and propose a tRansition rEconstruction framework for Compositional generAlization in next-POI prediction (RECAP). RECAP reconstructs long-tail transitions from two generalizable signals: multi-hop transitivity in the global transition graph and revisit evidence from a user's historical trajectory. It further uses warm-transition holdout training to discourage memorization of frequent transitions and encourage generalization from transferable signals. Experiments on multiple real-world datasets show that RECAP consistently improves prediction accuracy, with clear gains on tail transitions.

Problem

• human mobility prediction
• long-tail POIs
• transition-level sparsity
• compositional generalization
• next-POI prediction

Innovation

• compositional generalization
• transition-level sparsity
• long-tail POIs
• multi-hop transitivity
• warm-transition holdout training