Aspect-oriented Trust Based Mobile Recommender System

With the rapid advancement of wireless technologies and mobile devices, service recommendations have become a crucial and important research area in mobile computing. Although various recommender systems have been developed to help users to deal with information overload, few systems focus on personalized trustworthy recommendation generation for mobile users. In real life, trust plays an important role in the decisions related to sources that are used by human to take recommendations. This paper presents Aspect-Oriented Trust Based Mobile Recommender System (AOTMRS) that uses the concept of trust and Aspect Oriented Programming for advice-seeking and decision-making process similar to real life. The proposed system AOTMRS builds a mobility aspect and generates the trustworthy recommendations based on the user preferences and his demographic information such as location, time, need etc. AOTMRS exploits two peculiar characteristics of mobile information services such as “location-awareness”, i.e., the knowledge of the user’s physical position at a particular time and “ubiquity”, i.e., the ability to deliver the information and services to mobile users wherever they are, and whenever they need. Implementing user mobility in multi agent recommender system using conventional agent-oriented approach creates the problem of code scattering and code tangling. The mobility aspect separates the mobility crosscutting concerns, which in turn improves the system reusability, maintainability and removes the scattering and tangling problems of the recommender system. Moreover, because of resource limitations of mobile devices such as display size etc, AOTMRS uses personalized visual interface for mobile and computationally light algorithm at the user end in order to provide more effective and persuasive recommendations. The prototype of AOTMRS has been designed and developed for tourism recommendation system. Performance of the proposed system is evaluated using precision and recall metrics.