Big Data Analysis and Data Mining

The floating population is useful to figure out dynamic activities in urban area. Therefore, the prediction of floating population is required to practical use in urban and transportation planning. In Korea, the hourly floating population is estimated based on communication log of mobile phone. The communication log contains contents such as communication location, origin of mobile phone user. This paper is aimed to predict hourly floating population using mobile phone data, macroscopic data such as socio-economic index, and several data mining techniques. The data collected in Seoul, the capital of Korea, is used in this study. Also, the prediction accuracy by data mining technique is compared with each other, and the best model to predict hourly floating population is proposed in this study.