Location Based Context Aware Recommender System Through User-defined Rules

Abstract

Recommender systems are a subclass of information filtering system and are widely used in the e-commerce domain. They filter huge amount of data to provide personalized recommendations on services or products to users. Most of the existing approaches to develop a recommender system do not take into account contextual information such as weather, day, time, distance and location to provide recommendations. This thesis proposes a location based context aware recommender system through user defined rules that uses rules to provide context awareness in the system and a ranking function to generate top-k recommendations. The contextual data is defined by the users and is stored in the form of rules and RuleML is chosen as a rule based language. When an active user needs recommendations about nearby places then contextual data in the user-defined RuleML rules is extracted, evaluated, and top-k recommendations of nearby places based on the ranking function are presented to the user on the Google map.