Study On Key Techniques Of Web Mining For Intelligent Information Retrieval

With the rapid development of information on Internet, people find it becomes increasingly difficult to achieve the information that they need, the reason is that more and more traditional information retrieval methods can not meet the massive growth of information online, people even more look forward to the emergence of intelligent information retrieval to satisfy the growing information retrieval request.This dissertation researches some key techniques on Web mining for intelligent information retrieval.It mainly focuses on data preprocessing ,clustering of Web Pages or Web users.We improve some Web mining algorithms for intelligent information retrieval.An access intervals-based improvement was carried out of Session identification in web usage mining.The statistical result shows the page, access time accord with normal distribution .The access time threshold was adjusted by the web contents and site, structure on this condition.In the research on clustering of Web Pages and Web users,this dissertation analyze the common clustering methods and propose improved algorithms for K-means and DBSCAN.for K-means method,this dissertation gives a new method for selecting original clustering center based on data distribution to improve the clustering accuracy of K-means .For DBSCAN method,selecting the adjacency radius Eps and the number of adjacency data objects Minpts automatically is implemented.the improved method not only make the original clustering method more auto,but also extend the clustering ability.