Personalized recommendation framework based on CBR and CSP using ontology in a ubiquitous computing environment

摘要

This paper proposes OPERA (Ontology-based Personalized Information Recommendation) framework for ubiquitous users. To recommend personalized information in the ubiquitous computing environment, it utilizes case-based reasoning (CBR), constraints-satisfaction problem (CSP) solving techniques, and two kinds of ontology: context ontology and solution ontology. Context ontology can be used by CBR engine to find a solution or recommendation that considers past experiences of the user. So, the recommendation reflects the various contexts that imply the user's preferences. CSP adopts solution ontology to find the most appropriate alternative if the solution that is recommended by CBR is not available. The solution ontology is constructed by consideration on shared preferences for specific themes that are extracted from many unspecified Web users' experiences on the social network sites and blogs. We show the superiority of the OPERA framework through implementation of prototype and experiments.