Research On Book Recommendation Method In University Library Based On Reader Behavior Analysis

With the arrival of the era of big data,traditional document acquisition has been unable to meet the needs of readers.The existing ways of book recommendation include blindness,serious personal preference,low utilization rate and relatively passive nature.In view of the current development background of data mining,this paper proposes a Book Recommendation Model Based on reader behavior analysis.The construction process of the model is described in detail,in which the mining algorithm K-means needed in this paper is improved,the mining process of active books is discussed and the calculation method is given.The recommendation index of each book is calculated based on various factors.Finally,a book purchasing decision support system for purchasers is implemented.According to the application example,the research method is verified.The main work and results of this article are as follows:First,The mining process of active books is discussed.Taking the readers' behavior data of the University Library as the data source,the factors affecting the book activity are the number of borrowing times,the publishing house,the identity of the borrower,the number of retrieved times and so on.The calculation method is also given to make the book feature items quantized,which is convenient for subsequent cluster analysis and comparison.Second,Improved K-means algorithm.Aiming at the characteristics of active reader behavior data mining,we select the K-means algorithm in clustering algorithm.Because the clustering results of traditional algorithms are unstable and the accuracy is low,an improved K-means algorithm is proposed in this paper.Compared with other algorithms,the improved algorithm makes clustering results stable and the accuracy of clustering results is improved.Third,A method of book recommendation model is proposed.The model consists of active mining and matching books recommendation list consists of two processes.The active Book mining process based on the above;The book recommended by segmentation,single phase matching technology will be active and combined with existing books,gained recommendation list finally.Fourth,Implementation of a book recommendation and purchase decision support system.Based on the above book recommendation model method,combined with the correlation analysis and calculation method of recommendation index.Implementation of a book recommendation and purchase decision support system.It can enable the library to get the readers' needs voluntarily and greatly reduce the workload of the interview personnel.The book recommendation method based on reader behavior analysis is a useful supplement to the book recommendation mode.Model validation experiments prove the effectiveness of the proposed method,and meet the needs of University Library's book recommendation.