Behavioural Classification of Passengers in an Airport Terminal

Behavioural Classification of Passengers in an Airport Terminal

Rozema, L.M.

Hoogendoorn, S.P. (mentor)
Daamen, W. (mentor)
Gorte, B.G.H. (mentor)
Grosmann, R.J.F. (mentor)

Civil Engineering and Geosciences

Transport & Planning

Transport and Planning

2017-01-25

Given the observable characteristics of a passenger in an airport, is it possible to predict their behaviour? This thesis researches the possibilities for finding behavioural classes that can help predict passenger behaviour in a passenger demand forecast (PDF) system that is developed for the European PASSME project. The thesis introduces a framework that consists of two parts: clustering and classification. The clustering part of the framework finds behavioural classes, while the classification part determines how to classify passengers to such a class based on their characteristics. The classifier resulting from the trained framework can be used to classify new passengers into a behavioural class based on their observable characteristics. The framework has been tested on a survey data set, resulting in a performance of around 65%. Although this is not very high, the results can be regarded as promising due to the nature of the data that was used. As such, further development of the framework is likely to yield better results.

passenger behaviour
behavioural classification
Latent Class Analysis
cluster analysis
PASSME
Amsterdam Airport Schiphol
behavioural model
passenger demand forecast

http://resolver.tudelft.nl/uuid:c02e0b67-edbd-43ec-aa8a-a69d5004f1e3

52.309456, 4.7622836

Student theses

master thesis

(c) 2017 Rozema, L.M.