Prediction of Passenger Flow of Transit Buses Over a Period of Time Using Artificial Neural Network

Abstract

Neural network learning methods provide a robust approach for complex real-world data. The aim of this paper is to apply artificial neural network model for development of bus passenger flow prediction model. This model was developed to give real-time passenger flow of a particular bus over a period of time for transit agencies to build proactive strategies. The artificial neural networks (ANNs) are built with set of simple interconnected units where each unit takes real-valued inputs. For development of ANN model, the number of passengers at a given time (t) has taken as inputs and combinations of output play a vital role in achieving accurate predictive results. The purpose of this research is to develop and compare passenger flow prediction models which can provide accurate prediction in order to give real-time information. ARIMA and ANN models are considered and developed in this paper. These models can be used to implement advanced public transport system.