Web Search And Web Caching A Number Of Key Issues,

In the coming information age, the World Wide Web has been becoming an important platform, on which people publish and obtain their information.From the viewpoint of information consuming in the WWW, it is more and more important to effectively search the imformation in the environment of WWW. Topic distillation and exploration are two key techniques in this context. The first part of this dissertation focuses on the improvement of topic distillation and exploration with the aid of the linkage analysis. Based on the survey of current linkage analysis application on the Web search field, a similarity-based model on linkage analysis is proposed and an algorithm framework of topic distillation and exploration is developed based on the model.From the viewpoint of information supply in the WWW, reducing bandwidth and avoiding bottle-neck become key problems in the study of WWW applications. Web caching became the popular applied technique in such a context. The second part of this dissertation studies the Peer-to-Peer computing model and proposes a Web caching system based on the architecture of Peer-to-Peer.The main contributions of this dissertation are as follows:1. A survey of the linkage-analysis and its application on Web search is presented, and the future research directions in this field are also given.2. A similarity-based model on the linkage analysis is proposed. This model could be employed to observe the procedure of the linkage-based topic distillation. It provides the theoretical basis of improment for topic distillation.3. On the basis of the similarity model, the algorithm framework on the topic distillation and exploration is proposed, which can improve the quality of distillation using only the information ofhyperlinks. Furthermore, it also enables end-users to search less popular topics when multi-topics are involved in queries.4. On the basis of a self-configurable Peer-Peer computing model, a self-configurable web caching system is provided. The peer network can dynamically reconfigure itself based on the interests of its peers. Query will be routed from a peer to its neighbor that has the highest similarity value. The system adopts a self-adaptable hopping strategy...