A hybrid recommendation model for html5 mobile web applications

Abstract

In Recent years, the HTML5 Web technologies have become more and more popular. HTML5 provides a variety of new features and APIs that can be used by the applications running in terminals' browsers, especially on mobile devices. This new kind of application is called HTML5 Mobile Web Applications (Web Apps hereafter). Web Apps are written in simple Web languages such as HTML, CSS and JavaScript, which are transparent to the Web App Store or the Web Platform. As a result, the functions of a Web App can be easily analyzed. Based on this special characteristic, we propose a three- level hybrid recommendation model that can combine the Web Apps' implicit features (i.e. functions) as well as traditional explicit features (e.g., user ratings, user review) to build a probabilistic model between the user layer and the application layer. Consequently, this model can be used to recommend which applications are best suited for a user, as well as which users would be the potential consumers of an application. © Springer-Verlag Berlin Heidelberg 2012.