Alighting stop determination using two-step algorithms in bus transit systems

Abstract

Smart cards of most bus transit system only record boarding stops of passengers, which hinders direct exaction of alighting stops. In this paper, we associate multi-source information to derive transit OD matrix. Passengers are segmented into regular and irregular groups using K-means clustering based on trip frequency. In the first step, a deterministic model such as the trip chain analysis is respectively conducted on the two groups of passengers. For records that fail to be detected by deterministic rules, machine learning algorithms are developed to determine the alighting stop in the second step. All the destinations can be detected by the two-step algorithms, while only partial records can be determined in the past. Transfer trip recognition helps to distinguish transfer trips from single trips. This research develops a set of simple and applicable methodologies in alighting stop determination, which can be applicable more broadly to other transit datasets.