A Data Mining Approach to New Library Book Recommendations

摘要

In this paper, we propose a data mining approach to recommending new library books that have never been rated or borrowed by users. In our problem context, users are characterized by their demographic attributes, and concept hierarchies can be defined for some of these demographic attributes. Books are assigned to the base categories of a taxonomy. Our goal is therefore to identify the type of users interested in some specific type of books. We call such knowledge generalized profile association rules. In this paper, we propose a new definition of rule interestingness to prune away rules that are redundant and not useful in book recommendation. We have developed a new algorithm for efficiently discovering generalized profile association rules from a circulation database. It is noted that generalized profile association rules can be applied to other kinds of applications, including e-commerce.