Shared Bike Demand Prediction Based on Combined Deep Learnings

Abstract

The shared bike demand prediction can support shared bike scheduling activities and provide more convenient services for users. In this paper, a combined deep learning model, i.e., CNN-GRU-Attention model, is established. The model uses CNN network to extract local features of shared bike demand, GRU network to make predictions, and attention mechanism to extract important features. The parameters such as the number of neurons in the model are set experimentally. The simulation results show that the model has higher accuracy compared with other baseline models. It can fit the demand trend of shared bikes well and has good performance.