A new paradigm for personalized mashup recommendation based on dynamic contexts in mobile computing environments
Qingzhong Li, Ting Li, Guozhen Ren, Lizhen Cui, Wei He · Scientia Sinica Informationis · 2016
With the rapid development of mobile Internet and increasing number of intelligent terminals, Internet services are now integrated into people's daily lives. In mobile computing environments, challenges have emerged for unprofessional user oriented mashup servers with new features, such as dynamic contexts and user preference. In this paper, we focus on the entire construction process rather than individual service recommendation for mashups, and propose a context-based and user preference-based autonomous construction approach for mashups. The main idea is to construct a context- and preference-related conditional probability model for user behavior patterns based on pattern mining of historical mashup logs, which can provide support for the generation of optimal personalized mashup solutions and also recommend suitable web components. Analysis and simulation experiments conducted indicate that the proposed approach can autonomously generate personalized mashups with quality components without dependence on user professionalism.