Research On Long-tail Service Of Library Based On Data Mining

Today,the rapid development of Internet affects all walks of life.And the library service industry is the first to be affected.Electronic information publishers,together with Internet content providers are all competing in taking over the book information market actively.And also,they are trying to achieve the innovation mode of cooperation with information service providers such as information consultancy and search engine optimizer.Thus leads to the fade out of the viscosity of library users.And previous loyal readers have also turned themselves into the arms of the Internet big data resources.The development of the situation the Library facing is very serious.How to break shackles of traditional thinking mode,and taking advantages of the combination of library's traditional superiority with advanced technology will be the key pose to mend matters.Traditional library services tend to obey the Pareto's law,which means that it takes 20%of library resource to satisfy 80%of the readers.This thesis points out that Pareto's law is behindhand for library industry in consideration of Internet era and big date era.Libraries must try their best to satisfy their users so that users won't go away.The long tail theory,which is proved to be well performed in business area,can also be adopted to serve library industry.This thesis emphasizes the two roles in library service.One is the library resources,and the other is the user.There are three kinds of power which can speed up the progress of library's long tail.By analyzing the three kinds of power,we can conclude that it is applicable for library to transplant the long tail theory into its service.And the characteristics of the applicable conditions and the need for the long-tail adoption have also been demonstrated in this paper.Another original idea of this paper is to promote the idea of Two kinds of strategies of development long-tail service mode,which are meeting demands and creating demands.After a detailed study of the related theory of the long tail,data mining and text mining,this thesis has come up with a series of theory and methods which are suitable for library service.At the same time,this thesis has structured the data mining system on the Long-tail librarian service.The method of outlier data mining based on two times of clustering has also be proposed and validated by the research process of this paper.The results of empirical research are favorable.