DEVELOPMENT OF TRAVEL DEMAND MODEL USING MACHINE LEARNING TECHNIQUES

Abstract

As urbanization is happening at a sky-rocketing pace, the population in the Indian cities is growing rapidly which in turn is showing huge growth in travel demand. Motor vehicle ownership has risen alarmingly in several metropolitan areas over the past few decades at a pace of roughly 9% annually. The majority of Indians travel by bus, which not only meets their transportation needs but also serves as a required tool to draw in more affluent passengers (private car users). Unfortunately, there is a lot that can be done to enhance the state of bus services, including consistency, safety, and security. To draw discerning passengers to bus service and to prevent the expanding use of private vehicles, the quality of bus service urgently needs to be improved. User perception study is a practical and proven option to understand the need of travelers and to identify the reasons that deter the choice riders from using the public mode of transport. Revealed preference data is collected to understand the present condition and stated preference data is collected from the same respondents to understand their priority areas for improvement in the service quality for the future. The data is analyzed using conventional and advanced machine learning modeling techniques. Results indicated that choice riders are willing to shift towards the public mode of transport like ordinary and premium buses if certain attributes like travel information, security, comfort level, etc. are improved. These findings provide guidance for enhancing Kolkata's bus service based on the needs of choice riders. Even though the methodology was illustrated concerning the city of Kolkata, it might be used in other locations to derive metropolitan area service design and enhance bus services.