Concept lattices for visualizing and generating user profiles for context-aware service recommendations

Estimating user profiles for service recommendations in an efficient and manageable manner is very important. In a mobile setting, the service recommendation is more likely to depend on context. Since context data are multi-leveled, easy to reconfigure and dynamically changing, a sophisticated mechanism to handle these characteristics when mining for useful rules for ad hoc service recommendations from contextual data is necessary. Moreover, estimating the dynamic changes of the user profiles according to the change of context under a mobile setting is critical for context-aware service recommendations. However, legacy recommender systems having this capability are still quite rare. Hence, this paper aims to propose a user profile-discovering methodology. To do so, a concept lattice, which we will call the profile lattice, is considered for visualizing and generating user profile rules.