Data Mining Method for Smart Card Data Using Household Travel Survey: A Pilot Study of Public Transportation in Suwon, South Korea

Abstract

The aim of this study is to propose a new data mining method in order to conduct the data fusion 8 between household travel survey and smart card data. This study proposed the step-wise filtering 9 method, which categorizes a population group based on the activity patterns at first, then estimates the 10 temporal attributes of each activity from the household survey data. The estimated values are applied 11 for assigning trip purpose of the smart card data. A validation results suggest that the new method 12 shows a better performance than conventional method (simple filtering method) in the estimation of 13 trip purpose. This study finally verified the possibility the data fusion between two different data 14 resources by conducting a pilot study using smart card data in Suwon. For the future research, we plan 15 to advance the proposed method with probabilistic approach. That will be helpful for establishing a 16 public transit planning and related researches in transportation.