时间：2015年11月04日（星期三）上午09：00
地点：仓山校区成功楼603报告厅
主办：数学与计算机科学学院
主讲：武汉大学 谢榕教授
专家简介：Rong Xie received her PhD degree from University of Tokyo, Japan in 2003. From 2003 to 2007, she was a Science and Technology Promotion Special Researcher at the Center for Spatial Information Science, University of Tokyo. She was EC member of ISO/TC211 Group 6 (2005-2007), Member of CEOS WGISS (2004-2006) and Observer of ISO/TC211 Group 6 (2004). From 2007 to 2008, she worked as Technology Consultant at Japan Office, Intergraph Corporation, USA. Since 2008, she has been a professor at International School of Software, Wuhan University. Her main research interests include spatio-temporal data mining, artificial intelligence etc.
报告摘要：Mobile devices, such as mobile phones and other hand-held devices, are growing popularity. High-precision positioning technologies for determining geographical location of these devices are becoming increasingly available. Particularly, mobile phone data record people’s calling logs in everyday life, which reflecting our custom, pattern and lifestyle. It is opening a new way for us towards the study of our human activities by means of mobile phone location data. Our research aims at approaches to activity analysis from real mobile phone data records, including understanding individual behavioral patterns and characteristics, and also uncovering regularity of urban activities and their evolution following space and time. Our results show behavioral pattern mining, base station POIs extraction, user’s background identification and movement prediction at the individual level, as well as abnormal event discovery, downtown hot-period detection, spatial structure fitness analysis, and population activity patterns visualization at the city level, which help us more easily create a wide range of potential applications such as location-aware services/advertising, urban dynamics analysis, urban planning and social networking etc. New challenges are considered in our ongoing research when mobile phone data mining applications meet big data.