MFGCN: A Multimodal Fusion Graph Convolutional Network for Online Car-Hailing Demand Prediction

The rapid growth of online car hailing provides an excellent opportunity to provide convenient travel services. However, with the tremendous increase of users and online taxis, online car-hailing prediction systems face several challenges: 1) the difficulty of modeling nonlinear spatiotemporal interactions between users and vehicles, 2) the difficulty of incorporating context information and multimodal attribute enhancement data, and 3) the problems of data sparsity. To cope with these challenges, we propose a novel multimodal fusion graph convolutional network (MFGCN) for online car-hailing prediction. The model consists of a multimodal origin destination graph convolutional network module that contains three graph convolutional networks to extract spatial patterns from geography, semantics, and functional correlation; a multimodal attribute enhancement module that incorporates weather and temporal activity patterns; and a temporal attention skip-long short-term memory module that captures the periodic variations. Extensive experiments conducted on real-world taxi demand datasets show that MFGCN outperforms the state-of-the-art methods.