Abstract:

Addressing the problem of poor reader experience with relevancy sorting of traditional book retrieval systems, we propose the sorting algorithm based on the BM25 model and borrowing behavior prediction model. This algorithm synthesizes the relevancy algorithm of BM25 model and recommendation algorithm of borrowing prediction model, so that the result of sorting is subject to factors such as relevancy,cache time, borrowing condition, availability and collection venue. The result of tests shows that the comprehensive sorting algorithm makes the result of sorting more scientific and rational, and can enhance
reader satisfaction and rate of book borrowing.

Key words: BM25, Borrowing behavior prediction, Book retrieval, Sorting algorithm, Xunsearch