Estimating a nested latent class mode choice model on mobile network data

Angelica Andersson

doi:10.1080/03081060.2025.2593435

Estimating a nested latent class mode choice model on mobile network data

2025-12-01

0

PRE

AI

DOI：10.1080/03081060.2025.2593435

Original

Original request for help

Save

Abstract

En 中文

Declining response rates to travel surveys are an increasing problem for the estimation of accurate transport forecasting models. In this paper we investigate the use of mobile phone network data as a sole data source for estimation of a domestic long-distance mode choice model. We address two data-related challenges: the difficulties of estimating a model when bus and car trips are both observed as 'road' in the dataset, and distinguishing trip purpose. We successfully estimate a nested logit model with a nest that accounts for the differences in utility between bus and car, and we estimate a nested latent class model with the aim of identifying trip purposes. We find that while the proposed methods improve model fit, the latent class model cannot distinguish trip purposes clearly based on mobile phone network data alone. The paper thus demonstrates the benefits and limitations of mobile phone data for demand forecasting.

Keywords:

Demand model

mode choice

latent class

mobile phone network data

travel behaviour

long-distance travel

Estimating a nested latent class mode choice model on mobile network data

Abstract

Journal

Organization