Gong, Shi-Hao (Tongji University), Duan, Chuyu (Tongji University), Lan, Yiwen (Tongji University), TENG, JING (TONGJI UNIVERSITY), Guo, Jiayi (Tongji University)

Improving Multi-Modal Transportation Recommendation Systems through Spatio-Temporal Semantic Embedding

Scheduled for presentation during the Regular Session "S09b-Optimization for Multimodal and On-Demand Urban Mobility Systems" (WE-EA-T9), Wednesday, November 19, 2025, 13:50−14:10, Coolangata 3

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on October 19, 2025

Abstract

Multi-modal transportation recommendation plays a crucial role in Intelligent Transportation Systems (ITS). However, existing studies often provide homogeneous recommendations based on fixed objectives such as the shortest path, failing to fully model the complex spatio-temporal context and heterogeneous user travel preferences. To address these limitations, this paper proposes a spatio-temporal semantic embedding-based framework for multi-modal transportation recommendation. The framework leverages a pre-trained language model to semantically encode heterogeneous spatio-temporal features and candidate travel plans and employs a Transformer-based context-aware layer to capture deep interactions among features. The resulting spatio-temporal semantic embeddings are then fused with structured user attributes and fed into a multi-layer perceptron (MLP) model for travel mode prediction. Experiments conducted on over 300,000 real-world navigation records from Beijing demonstrate that the proposed method outperforms five state-of-the-art baseline models across multiple metrics. The results validate the potential and applicability of spatio-temporal semantic embeddings in transportation mode recommendation and suggest a promising shift for ITS from static path planning towards cognition-enhanced decision support paradigms.