Prediction of Modal Shift Using Artificial Neural Networks

Various public transport concepts have been developed to provide solutions to the ever growing problem of traffic in modern times. For instance, intelligent subscription bus service is one of them. This concept aims to provide a means of transport at near private car comfort as well as at near public transport cost. By this means, a shift from other modes of transport, especially private car, to public transport is aimed to encourage use of public transport. An artificial neural network model has been developed in this study to be able to calculate modal shift using three sources of data obtained from two questionnaire surveys conducted at Akdeniz University campus and a computer model's output (based on shortest route algorithms). The relationship between the results of the second questionnaire survey and the other data have been entered into Weka and Rapid Miner programs, the accuracy of this machine learning has been calculated and finally the modal shift originated by the intelligent subscription bus services has been estimated. The findings have yielded very reliable results which revealed the potential of applying the technique easily to similar problems.