The Research On Adaptive Personalized Recommendation Algorithm Of Book

With the rapid development of information technology,the amount of data accumulated in the rapid growth of people,how to extract useful knowledge become a top priority from the mass of data.Library Business background under the big data will be faced with complex,diverse content data as an object to dig deep content for the purpose of the specialized requirements.Libraries need not only structured data reader now understand what services,but need to use large amounts of unstructured,semi-structured data to make multi-dimensional analysis,the reader's perception of the real needs,to provide readers with personalized service.Appear personalized recommendation technology for library books recommended service provides a good solution ideas.In this paper,an adaptive personalized book recommendation algorithm research objectives,the flow of business data to the library for the study,combined with collaborative filtering,context-aware recommendation,user behavior analysis processing characteristics and other related technologies,research and analysis solution library individuation the method of service,and experiments show the feasibility and effectiveness of the book recommendation algorithm.Research work of this paper includes the following aspects:(1)Improve readers' interested model presenting method."Zhong Tu Fa" books classification criteria for classifying books,built on words of the book "reader-Books types" access matrix can be an effective solution based library resources,"the reader-Books' access matrix due READERS' BORROWING with respect to the number of library books too few,and many readers did not produce the same behavior borrow books from each other,due to the sparsity of the matrix is too large.(2)bonding time contextual information,the relationship between mining Library Circulation time information and business data reader interest change,according to the reader has time library behavior and by the length of the reader to borrow books acts set different weights,can make readers interested in the model to reflect the real needs of readers,book recommendations to improve the effect.(3)Establish analytical method based on readers' negative feedback behavior.combined with business data library circulation records Readers negative feedback acts readers behavioral characteristics process,negative feedback on negative samples collected behavior,improve the reader interested in the model,which identify the target audience of "nearest neighbors",predicting the target level of interest on the target reader of books,improving personalized recommendation accuracy.(4)Create a self-adjusted personalized book recommended method to the public library.To test the plausibility and effectiveness of this method,five controlled experiments were designed and carried out.