Influence Factors on Travel Mode Preference of Working Residents Living Far Away from Downtown Area on Workdays: A Hybrid Method Integrating Correlated Random Parameter Logit Model and Apriori Algorithm

Abstract

This study presents a detailed analysis on the characteristics of travel mode preference of working residents living far away from downtown area on workdays, using GPS-based activity travel diary data from Shangdi area (Beijing). A hybrid method integrating correlated random parameters logit (CRPL) model and Apriori algorithm is put forward to explore the influence factors affecting travel mode preference. First, the CRPL model is established to explore significant factors and capture the unobserved heterogeneity and correlations among the unobserved characteristics on the travel mode preference. Then, these significant factors are used to generate association rules by Apriori algorithm to investigate statistical associations between the specific travel mode preference and these significant factors. Ten significant factors are found in the CRPL model, in which three factors including daily time spent on maintenance activity, household has elderly or not, and distance between house and shopping mall are normally distributed. The results of the Apriori algorithm indicate that some factors combined with other factors could significantly influence working residents’ travel mode preference. For example, the combination of lower annual household income and shorter distance between workplace and the nearest bus stop is highly associated with green travel mode preference. Moreover, the results show that the proposed hybrid method not only demonstrates the consistency of the results of the two methods, but also plays a complementary role in exploring more information on travel mode preference. This research hopes to give regulators a better understanding on how working residents living far away from downtown area choose their travel mode, so as to develop more effective and targeted measures for reducing private car use and alleviating workday traffic congestion.

Keywords: travel mode preference, correlated random parameter logit model, Apriori algorithm, working residents, green travel mode