No próximo dia 2 de Outubro de 2019, pelas 14h00 no Anf. 2 (FC6-0.29) do DCC, Masahiro Kimura irá dar uma palestra intitulada "Extracting Influence Structure in Geographical Attention Dynamics".
A palestra é organizada pelo DCC-FCUP e pelo grupo de investigação LIAAD-INESCTEC e é aberta a todos os interessados.
Short Bio
Masahiro Kimura received his B.S., M.S, and Ph.D. degrees in mathematics from Osaka University, Osaka, Japan, in 1987, 1989, and 2000, respectively. In April 1989, he joined Nippon Telegraph and Telephone (NTT) Corporation, Tokyo, Japan, and mainly worked at NTT Human Interface Laboratories and NTT Communication Science Laboratories. In March 2005, he left NTT. In April 2005, he joined Ryukoku University, Kyoto, Japan. Currently, he is a professor in the Faculty of Science and Technology. His research interests include complex network science, data mining, and machine learning.
Title
"Extracting Influence Structure in Geographical Attention Dynamics"
Abstract
With the development of smart mobile devices, location acquisition technologies and social media, a large amount of event data with spatio-temporal information has become available and offers an opportunity to better understand people's location preferences and mobility patterns in a sightseeing city. We address the problem of modeling geographical attention dynamics for a sightseeing city, that is, modeling the occurrence process of POI (point-of-interest) visit events in the city in the setting of a continuous time-axis and a continuous spatial domain. We propose a probabilistic model for discovering the spatio-temporal influence structure among major sightseeing areas, and aim to accurately predict POI visit events in the near future. The proposed model is constructed by combining a Hawkes process with a time-varying Gaussian mixture model in a novel way and incorporating the influence structure depending on time slots as well. We develop a method of inferring the parameters in the proposed model from the observed sequence of POI visit events, and provide an analysis method for the geographical attention dynamics. Using real data of POI visit events in Kyoto, we demonstrate that the proposed model significantly outperforms the conventional models in terms of predictive accuracy, and reveal the spatio-temporal influence structure among major sightseeing areas in Kyoto from the perspective of geographical attention dynamics.