Improving performance of a mobile personalized recommendation engine using multithreading

Abstract

Popularity of social networking services (SNS) and location-based SNS (LBSNS) have an influence on lifestyles of many people. Furthermore, the advancement of mobile technology enables people to share their interests and lifestyles to their friends conveniently. These factors cause the Internet to become massive personal information resource. A mobile personalized recommendation (MPR) engine plays an important role in offering solely essential information to prevent information overload for the users. Unfortunately, processing time of traditional MPR engine is high. This paper proposes an approach to improve performance of MPR using multithread programming (MP). The experimental results indicate that the multithread programming (MP) could deliver higher performance than sequential programming (SP), especially speedup between 5 and 7 times approximately.