Data mining Technology-Based Design and Implementation of University Library Personalized Service

Abstract

With IT and network technology becoming more and more mature, the application of database technology is deeper, university library data storage and management also significantly increase during this process, how to expand the space and scope of university library services so that readers can access information resources they need is put on the agenda, university library of personalized service is born. In order to make university library personalized service able to have a strong theoretical work as a basis, the paper bases on data mining technology, focuses on cluster analysis and association rules models, analyzes and researches the library personalized service. Introduction In today’s era of the rapid development of information and digital technology, the management of university libraries data is increasingly complex and diverse, in order to meet the needs of the times, management of university libraries also tends to digital and information, while the data mining technology is precisely to solve this problem, providing a good platform for readers to quickly search out the needed date in such a large amount of library data. Through cluster analysis and association rules model, data mining technology transforms the traditional library services to personalized and humanized services, to facilitate the reader to retrieve and read. Cluster analysis According to some rule,divide a large amount of clutter data stored in database into a plurality of data sets, the data in each data set are not the same, but there is the similarity of the date in the same data set. Such clustering analysis on the establishment of a macro concept, readers in search of information you can consult the relevant literature to more relevant content, convenience of the reader effectively summarized. The association rule model The value associated with the name suggests is there are some laws between two or more variables on the known presence of these variables are associated. Association rules between the large amount of data is from the collection of items to find meaningful association, thus achieving the objective laws of technical methods of understanding of things. Definition: Set I = {i1, i2 ..., im} is the set of all projects, set A is a collection composed by the project, called item-sets. Transaction T is a subset of items, each transaction has a unique transaction identifier Tid. Transaction T contains item-set A, if and only if AT. If the item set A contains k items, claimed it as k item set. D is a transaction database, the number of items appearing in the set A transaction database D, D as a percentage of the total transaction is called term support sets (support). If the support of the item set exceeds a user-specified minimum support threshold, the set is said to frequent item sets (or large items). Support refers to the frequency pattern appears in the rule, if the transaction database has s% of transactions contain XY, XY called association rules in D, the degree of support for the s%, in fact, can be expressed as the probability P (XY), That support (XY) = P (XY). Trust refers to the