Multi-step Point-of-Interest-level Crowd Flow Prediction Based on Meta Learning

Point-of-Interest-level(POI) crowd flow prediction is an important task for businesses and consumers. Based on POI-level crowd flow prediction, businesses could make more reasonable business arrangements, and consumers could make more suitable travel plans. However, POI-level crowd flow prediction is a challenging task for two aspects: 1) Compared with region-level crowd flow, the area of POI is smaller and the fluctuation of POI-level crowd flow is greater; 2) There are diverse temporal correlations of different POIs and varies over time. To tackle the above challenges, following the antoencoder architecture, we propose a multi-step POI-level crowd flow prediction model(Ms-PLCFP) to predict the crowd flow at all POIs at once. A meta learner is used to obtain meta knowledge from POI category, POI popularity, etc. Then meta-RNN+ is applied to model the relations between temporal correlations and meta knowledge so as to capture diverse temporal correlations. Furthermore, a multi-scale temporal attention mechanism which contains multiple different scales of temporal attention is employed to smooth input crowd flow at lower level and capture global dependencies of input crowd flow at higher level. We evaluated Ms-PLCFP on two real-world datasets and Ms-PLCFP achieved significant improvements over the baselines, which shows the effectiveness of our model.