ABSTRACT
A B S T R A C T The features of smart card data, i.e., precision, continuity and long-term observation enabled us to analyse the dynamic characteristics of travel behaviour. In order to explore the dynamic characteristics of a large amount of information in the data set, previous studies have developed data mining methodologies and applied them. However, the smart card systems were not specialized to collect a data set for travel behavioural analysis. The smart card data offer only fragmentary information on travel behaviour though they can provide accurate and continuous long-term data, which is difficult to achieve via conventional behavioural surveys. In order to supplement absent behavioural attributes in the smart card data, this study proposes a data fusion method of smart card data with the person trip survey data. The results of the data fusion enable us to analyse the continuous long-term features of the trip purpose of transport users, which are difficult to get from either the survey-based data or the smart card data. These results enable us to know specific behavioural segments, which caused changes in travel demand.
