Forecasting travel behaviour from crowdsourced data with machine learning based model

Information and communication technologies have become integral part of our everyday lives. It seems as logical consequence that smart city concept is trying to explore the role of integrated information and communication approach in managing city’s assets and in providing better quality of life to its citizens. Provision of better quality of life relies on improved management of city’s systems (e.g., transport system) but also on provision of timely and relevant information to its citizens in order to support them in making more informed decisions. To ensure this, use of forecasting models is needed. In this paper, we develop support vector machine based model with aim to predict future mobility behavior from crowdsourced data. The crowdsourced data are collected based on dedicated smartphone app that tracks mobility behavior. Use of such forecasting model can facilitate management of smart city’s mobility system but also ensures timely provision of relevant pre-travel information to its citizens.