Data Mining Algorithms In Bibliographic Recommendation System

With the rapid development of science and technology, the digital library has become a kind of digital information organization.Information Service is the main one in Digital Library, in which personalized information service has become the mainstream mode. In recent years, how to extract the wanted ones from the mass information according to the users'reading habits has become a hot research in the library industry.Data mining is developing rapidly in recent years. It is a technology to discover and identify potential, unknown and valuable information. So applying it to library Automation system can really meet the requirement of personalized information services in libraries.Aiming at the fact that Library Automation System does not have data mining capabilities, this paper applis the association rule in Data Mining to the Henan normal university library automation system based on the large amount of readers'borrowing log. Thus, a detail method which provides personalized recommendation service for readers is found in this paper and this method meets the readers'need well.In order to make the recommended books more close to the readers'taste and provide an intuitive choice for readers, an effective algorithm FP_MAX is put forward and a personalized recommendation system which is suitable for the readers' borrowing habits is designed with the improved FP_MAX in this paper. For the purpose of providing a higher active FP_MAX algorithm, the unnecessary sets of patterns are reduced according to the nature of Maximal frequent patterns and an intermediate result is set to decrease the scope of inspection when the Maximal frequent pattern is verified.