DiffMove: Group Mobility Tendency Enhanced Trajectory Recovery via Diffusion Model

TL;DR

DiffMove tackles trajectory recovery under sparse data by fusing population-level group mobility tendencies with individual mobility preferences within a diffusion-model framework. It builds a group tendency graph and embeds it into location representations, while simultaneously extracting historical and current personal preferences and refining them through a diffusion-based mobility distribution. The approach yields consistent improvements over a wide range of baselines on two real-world datasets, demonstrating robustness to high data sparsity and strong generalization across domains. This work offers a practical, scalable method for denser, more accurate mobility trajectories with potential benefits for urban planning, routing, and epidemiological modeling.

Abstract

In the real world, trajectory data is often sparse and incomplete due to low collection frequencies or limited device coverage. Trajectory recovery aims to recover these missing trajectory points, making the trajectories denser and more complete. However, this task faces two key challenges: 1) The excessive sparsity of individual trajectories makes it difficult to effectively leverage historical information for recovery; 2) Sparse trajectories make it harder to capture complex individual mobility preferences. To address these challenges, we propose a novel method called DiffMove. Firstly, we harness crowd wisdom for trajectory recovery. Specifically, we construct a group tendency graph using the collective trajectories of all users and then integrate the group mobility trends into the location representations via graph embedding. This solves the challenge of sparse trajectories being unable to rely on individual historical trajectories for recovery. Secondly, we capture individual mobility preferences from both historical and current perspectives. Finally, we integrate group mobility tendencies and individual preferences into the spatiotemporal distribution of the trajectory to recover high-quality trajectories. Extensive experiments on two real-world datasets demonstrate that DiffMove outperforms existing state-of-the-art methods. Further analysis validates the robustness of our method.

Figures (5)

• Figure 1: The framework of DiffMove.
• Figure 2: The architecture of denoising network $\mathcal{F}_{\theta}$.
• Figure 3: The sampling process of DiffMove.
• Figure 4: Impact of each component on two datasets.
• Figure 5: Sensitivity of hyper-parameters.