The Research Of Personalized Recommendation Algorithm Of Library Books

The rapid development of Internet technology brought the problem of information overload to all areas and industry, including library. In the era of the past, data is so poor that book recommendation service are often reply on human recommendation or even not exist. However, due to the different borrowing needs by every reader, the way to service is hardly to meet the vast majority of readers' needs.Based on creating a binary relationship between reader and book and the existing borrowing records or similar relationship, personalized recommendation system mining borrowing needs for each reader,who own unique backgrounds.The nature of this process is information filtering,which means recommend some specific books to its readers.This paper introduces the basic principles of collaborative filtering algorithms, and specific methods which collaborative filtering algorithms based on users and books provide in the University Library Collection of books recommendation area. Meanwhile,to mine the practice law and borrow law of college library,we analyse the existing borrowing records in the aspect of timing law and the utilization ratio of library books in this article. In the third chapter, we study the collection of books recommendation algorithm based on readers properties and build such a readers_featured model based on readers' background, on which similarity between readers is calculated based. In the final experiment, we receive a better result in terms of execution or performance of the algorithm than collaborative filtering.