A Novel Neural Network Model for Demand Prediction of Bike-Sharing
説明
Accurate demand prediction of bike-sharing is a prerequisite to reduce the cost of scheduling and improve the users’ satisfaction. However, it is very difficult to make the prediction absolutely accurate due to the stochasticity and nonlinearity in the bike-sharing system. In this paper, a model called pseudo-double hidden layer feedforward neural network is proposed to approximatively predict the practical demand of bike-sharing. In this neural network, an algorithm called improved particle swarm optimization in extreme learning machine is proposed to define its learning rule. On the basis of fully mining the massive operational data of “Shedd Aquarium” bike-sharing station in Chicago (USA), the demand of this station is predicted by the model proposed in this paper.
