Book recommendation systems can benefit commercial websites, social media sites, and digital libraries, to name a few, by alleviating the knowledge acquisition process of users who look for books that are appealing to them. Even though existing book recommenders, which are based on either collaborative filtering, text content, or the hybrid approach, aid users in locating books (among the millions available), their recommendations are not personalized enough to meet users' expectations due to their collective assumption on group preference and/or exact content matching, which is a failure. To address this problem, we have developed PBRecS, a book recommendation system that is based on social interactions and personal interests to suggest books appealing to users. PBRecS relies on the friendships established on a social networking site, such as LibraryThing, to generate more personalized suggestions by including in the recommendations solely books that belong to a user's friends who share common interests with the user, in addition to applying word-correlation factors for partially matching book tags to disclose books similar in contents. The conducted empirical study on data extracted from LibraryThing has verified (i) the effectiveness of PBRecS using social-media data to improve the quality of book recommendations and (ii) that PBRecS outperforms the recommenders employed by Amazon and LibraryThing.
展开▼
