Design And Implementation Of Library Management System Based On Hybrid Collaborative Filtering Recommendation

With the rapid development of information technology, books storage and management model in traditional library should be constantly improved according to the need of users. At present, as the category of related information and the amount of information in the library is becoming more various and huge, it has been more desirable for the intelligent processing on the information. With the readers more and more diversified need for the knowledge and information, individualized service in the library can be regarded as the main way to further meet their need. Meanwhile, there is the mutual relationship between different books, which can be used as the basis for the individualized service. Considering most libraries having not setting up effective individualized services at present, better services can not be provided for readers. Therefore, through the data mining analysis on the information of searching books, keeping books, evaluating books, in the process of library management, collaborative filtering technology is introduced to recommend books for users and better meet their needs.Firstly, based on the practical situation of books management in library, this paper analyzes the applied requirement of books management system, and offers an analysis of systematic business process and of functional definition as well as its performance need. In terms of business process, four main business processes such as keeping books, the user management, books management and library management are elaborately analyzed. In the aspect of requirement analysis on system function, it is functionally defined by five types of personnel: the readers, the library staff, library manager, system administrators, and staff external system. In terms of performance, the actual requirements and design ideas of system operation on stability and maintainability, the corresponding speed and the security are provided.Secondly, on the basis of the requirement analysis, books management system is designed based on collaborative filtering from both users and projects. A design method facing to the object and the modular is adopted by the system according to the thought of software engineering, and the system function structure design, logic framework design and data structure design are all offered in this paper. Considering the collaborative recommendation based on the user and the system recommendation based on the project, the accuracy of the recommendation results to some extent can be improved. This paper proposes a filtering algorithm of hybrid system based on the fusion of the user and the project. The algorithm can realize individualized recommendation of books through the calculation of similarity, similar neighbor selection, collaborative forecasting and fusion calculation. Validity and reliability of the proposed algorithm are testified through a set of experiments.Finally, combined with design specifications and system design procedure, this research realized a books management system with a friendly interface, convenient operation, and individual recommendation. By introducing a collaborative filtering recommendation technology into the books management system, mining the information of users’ history evaluation and providing users with individualized recommendation, there is important practical value in improving the library’s service quality and users’ satisfaction.