A design of context-aware framework for conditional preferences of group of users

Abstract

Due to the dependency of users’ preferences, which change over time, there is a need to generate a recommender framework that can handle users’ conditional preferences. Since the most existing context-aware frameworks have been created for individual users, until now, little academic attention has been paid to the conditional users’ preferences in group context-aware recommender systems. In dynamic domains, users’ preferences are continuously affected by the domain entities, that is any object acting as service providers (e.g., restaurant, weather, users, etc. could be the entities in a restaurant selection guide). Hence, in this study we propose a context-aware framework that provides service(s) according to the current context of entities and the current users’ preferences, which are naturally conditional. In addition, our framework is self-adaptive because it can adjust its own behavior (that is, the service it provides) based on the current context of entities. The Hyperspace Analogue to Context (HAC) model is used to abstract and represent the multi-dimensional entities’ context to the system. The main goal of the proposed framework is to keep up at high level of satisfaction degree of group of users in a dynamic domain. We conclude this paper by simulating and evaluating our framework on a concrete scenario.