Adaptive Clustering Algorithm And Its Application Technology Research And Implementation

With the rapid development of network communication techniques and the widely application of the database technology, modern enterprises are more likely to do business or publish information through the Web. E-commerce systems have produced a mass of data as they serve customers. Hidden knowledge or patterns could be discovered by the enterprise administrators through mining operations on these massive datasets, and applied to gain decision advantages over competitors. This has become one of the main applications of data mining techniques.Under this background, we firstly expatiated the classification, primary aspects, and applications in e-commerce of data mining. After we discussed popular clustering techniques in detail, an improved clustering algorithm, Adaptive K-means, is presented on the basis of analyzing the advantages and disadvantages of the traditional K-means. The new algorithm is superior to the traditional one that it can determine the cluster number and the initial centroids adaptively, and reduce the blindness and subjectivity to some extent. Simulation experiments are performed to testify the correctness and validity.As a pre-research of the "ZHENXIANG medical e-commerce system II", our work must be compatible with the existing applications, in order to be integrated as a component. A data mining system, Weak-on-Web, is presented and implemented in Browser/Server style. It supports the Servlet 2.3/JSP 1.2 specifications and the J2EE architecture, allowing the data analysts to work through a Web browser from almost anywhere. Adaptive K-Means is also implemented and verified in this prototype, and encouraging results are made.Finally, the main contributes of this thesis are summarized and some future works are outlined.