Mobile User Modeling for Intelligent Recommendation (2016-2018)
Abstract
This project seeks to develop a mobile user modelling framework which accurately infers mobile users' location-time-dependent interests and spatial mobility patterns from their daily activity records and social ties in geo-social networks. It will then build an intelligent recommender system to provide location-based, real-time and personalized services by effective fusion of the multi-faceted knowledge about the mobile user learnt in the modelling framework. The proposed research will address multiple challenges arising from data sparsity, data heterogeneity, data incompleteness and efficiency issues in processing large amounts of data, and the resulting techniques apply to various location-based services, mobile advertising and marketing.
