Research On The Application Of Data Mining In The Personalized Service Of College Libraries

With the rapid development of the information technology, the appearance andpopular of lib2.0and informationization.College library has accumulated large amountsof data in the process, So that the reader can’t fast and convenient get that they needinformation resources. The reasons for college library necessary expand service spaces.So the individualized service became the mainstream mode of the library service, Thismodel transitions the passive service of the traditional library into active service,recommending books to users depending on their interest profile actively.First this paper comparative analysis both of present situation of the application ofpersonalized service system of domestic and international university library. Systemicexplaination about the concept of data mining、task、method and process, in themeanwhile, introduces about related knowledge of the book classification and digitallibrary individualized information service mode. This paper has a discovery withanalysis of Library integrated system, which used by majority of college library. Thediscovery is that these systems don’t have data mining feature. In this paper, we willdesign a supporting system of based on data mining technology personalizedrecommendation based on the library-integrated system. We make a detailed illustrationto the specific module composition, functional description and working flow from angleof fundamental research.In the study of concrete realization, this paper using data mining technology,making the log of borrowing and reading of Chong Qing university library as theresearch object, studying concrete realization of data processing (Including dataacquisition) and mining analysis in the personalized recommendation system of collegelibrary, Further realize the information behavior and the establishment of the knowledgebase and put forward some concrete results mining recommendations. We mainlyanalysis for mining from two aspects.On the one hand, make effective group classification of readers of library usingclustering analysis and each group that has similar borrowing mode. According to theresults of the clustering build user model.On the other hand, With all the readers and are already clustering reader for miningobject use of association rules mining for mining analysis. Find out the strong relevancebooks and extracting rules model, establish knowledge base. Finally, volunteer actively recommend books according to the rules that eventually mode matching with readersFinally, This paper make a comprehensive summary for this research, and look tothe future direction of further study.