Understanding and Modeling the Social Preferences for Riders in Rideshare Matching

Yu Cui, Ramandeep Singh Manjeet Singh Makhija, Roger B. Chen, Qing He, Alireza Khani

Research output: Contribution to journalArticlepeer-review

Abstract

Ridesharing is the sharing of trip segments from one place to another among multiple travelers, obviating others’ needs to drive themselves. By having more than one occupant sharing a vehicle, ridesharing aims to reduce personal resources and costs, such as fuel and trip-related costs, and driver stress. The objective of this paper is to model the social preferences of rideshare passengers. We identify challenges and barriers people face in ridesharing with respect to whom they share the ride with and model these social preferences to determine the probability of matching for rideshare demand forecasting. An online survey instrument was designed and distributed among the people residing in the United States to uncover their preferences for ridesharing, in addition to the attributes of potential rideshare passengers. Furthermore, using the survey data, a discrete choice model with latent variables was estimated to uncover the relationship between social preferences and matching. We identified 13 attitudinal dimensions characterizing social preference from the survey responses. These 13 variables were further distilled into four latent variables using factor analysis. Four models were estimated for each latent dimension to predict the probabilities of a person pleasantly experiencing his/her shared rides in social aspects from his/her attributes and preferences. Based on the estimated choice model, we developed a matching index derived from preference probabilities that give a compatibility ratio between riders.

Original languageEnglish (US)
JournalTransportation
DOIs
StateAccepted/In press - 2020

Keywords

  • Factor analysis
  • Latent variables
  • Ordinal logistic regression
  • Ridesharing behavior
  • Social preferences
  • Survey

Fingerprint Dive into the research topics of 'Understanding and Modeling the Social Preferences for Riders in Rideshare Matching'. Together they form a unique fingerprint.

Cite this