Public Bicycle Prediction Based on Generalized Regression Neural Network

Abstract

Nowadays building a green and efficient public transportation system for the expanding urban population is undoubtedly a big challenge. In recent years, public bicycle system has been widely appreciated and researched worldwide. Unlike traditional public transportation system, public bicycle system doesn't need to follow fixed schedule. This flexibility brings high efficiency as well as uncertainty-we don't know whether there are available bikes or bike stands when they are indeed needed. This paper aims to predict the number of available bikes at given future time point so as to optimize the user's travel choices. In this article, we propose a new prediction model and use generalized regression neural network as our prediction algorithm to optimize prediction accuracy. Experimental results show that in this way we can properly handle this nonlinear problem.