Abstract:

Modern libraries provide only modest book recommendation services. We present in this paper a novel book evaluation and recommendation approach based on collective intelligence. Inspired by PageRank, this approach evaluates books based on their mutual recommendation relations provided by online bookstores’ massive user group, before recommending books from not only one particular category but also relevant categories. The experimental results show that it can provide more comprehensive recommendation which pinpoints a user’s expectation.

Key words: collective intelligence;book evaluation;BookRank;recommendation