Data Mining Applications, Personalized Services In Digital Libraries

To solve the problems in the traditional personalized service system of digital library, such as deficiency of data analysis, phenomena of information alone islands, and singleness of personalized services modal, etc, the paper, introducing data warehouse (DW), on-line analytical processing (OLAP) and data mining (DM) concepts together with network technology, carries through a comprehensive theoretical and empirical study on digital library.The paper, with DW, OLAP and DM techniques adopted, has analyzed and delved into the data of reader, resources, and the interviews between them, and then designed and realized an individual service system of university digital library based on the association analysis and reader interest models. The system analyzes the data and identifies readers' interests before providing them with personalized services. The main contents of this dissertation are summarized as follows:(1) Adopt a new Decision Support System (DSY) to design the model of the personalized service system, which emphasizes analyzing resources and identifying readers' interests, in a digital library.(2) Adopt a dimension-modeling method to design the data warehouse logical mode based on such three subjects as reader analysis, resource analysis and interview analysis.(3) Put forward an enhanced algorithm, which associates AprioriTid with transaction reduction and item reduction technique. In this algorithm, candidate set generation and the support calculation of each item set is created after each transaction is compressed and connected. As the candidate set adopts key words to identify in this algorithm, thus the process of pruning and string pattern matching is removed from AprioriTid algorithm. It is shown, from the results of tests, that the algorithm outperforms AprioriTid algorithm.(4) Establish a multidimensional data cube of information and use the AprioriTid enhanced algorithm to analyze document resources for analyzing resource interviews and the association between readers and resources.(5) Design and realize a reader interest model based on a resource classification tree, which, through active and passive interaction with a reader, constantly receives and speculates the reader's interest, and accumulates the reader's information requirement bias, thus realizing self-adaptive personalized service in digital library.The research work in this dissertation has laid a good foundation for the construction and development of the personalized service system of a digital library.