Inference of Bus Get-off Station Based on Decision Tree C4.5 Algorithm

With the urbanization and the expansion of urban scale, the rapid growth of the public travel demand is increasing. Public transport system has collected a large number of passenger travel data and these data contain a large number of passenger travel information. In order to obtain the travel characteristics of bus passengers, a bus drop-off station inference algorithm based on the decision tree C4.5 algorithm of machine learning is proposed. In this algorithm, the passengers are classified into different categories according to their travel data. These travel data are also be described using a set of attributes. Through supervised learning, the mapping relationship between attribute value and category is found. Based on the relationship, the off-site information of new passengers is inferred according to the new travel data. Taking Tianjin bus transport system as an example, the proposed algorithm is used to infer the drop-off station of passenger. The inference result shows that at least 75% drop-off stations can be obtained using our algorithm, which can provide a useful support for further study of passenger travel behavior, bus schedule and traffic control.