Journal of Choice Modelling

This code allows parallel estimation and parallel prediction of kernel Multinomial Logit Model. In this code, we use a subset (5000 observations) of the data used in this paper.

Data, Matlab codes and results of the simulations and models presented in the paper.

This dataset was part of a master thesis analysing the use of different decision rules in best-worst (BW) choice models. BW models require respondents to keep making choices within a choice set (removing the selected alternative each time) until no alternatives remain and a ranked order is implicitly obtained. The case for the survey was a park and ride facility choice, where respondents were given five alternative park and ride facilities, with different access times, prices, travel times by public transport, modes of public transport and operating frequencies. Responses were obtained online, in June 2018. In total, 108 valid responses were obtained.

This dataset was part of a master thesis analysing the use of different decision rules in best-worst (BW) choice models. BW models require respondents to keep making choices within a choice set (removing the selected alternative each time) until no alternatives remain and a ranked order is implicitly obtained. The case for the survey was a park and ride facility choice, where respondents were given five alternative park and ride facilities, with different access times, prices, travel times by public transport, modes of public transport and operating frequencies. Responses were obtained online, in June 2018. In total, 108 valid responses were obtained.