Study On The Application Of Automatic Focused Searching

Along with the boom of the Web, information service institutions and their end-users begin to focus on how to make full use of Web resource effectively and at low-costly. The aim of this dissertation is to exploit the automatic construction methods of Web subject resource according to the study on the application of automatic focused searching. Automatic focused searching can automatically retrieve the free Web resource, get rid of the dependence on experts, reduce the cost of construction, and improve the speed, efficiency, and quality of Web subject resource construction.The major research work and contributions of this dissertation are as follows:(1) The basic theory of focused searching and the construction modes of Web subject resources are investigated respectively. Based on these investigations, the thesis explores the related techniques of automatic focused searching and brings forward a functionality framework for automatic construction of Web resource.(2) Based on DFSA algorithm (Deterministic Finite State Automaton) and combined with Quick Search algorithm, this paper analyses and implements a new multi-pattern string match algorithm that consumes less half of memory space of standard DFSA algorithm. This Algorithm can be used to speed up the processes of feature extraction and classification of Web pages.(3) From the viewpoints of sociology, bibliometrics and computer science, this thesis analyses basic theory of Web hyperlink, and then implements a new technique that can be used to discover new related resources on the Web using the modification of the classical HITS algorithm (Hypertext Induced Topic Search). Based on the bibliographic co-citation and coupling, the paper implements some algorithms finding related Web page, analyses their performance in comparison with the related technique of Google and Alexa Internet.(4) This dissertation analyses and designs a collaborative focused crawler model. Because of using Web link analysis technique and tunneling technique, the crawler can significantly improve its topic coverage and topic precision, overcome the crawler's inherent disadvantage to some degree that the topic coverage and topic precision seriously depends on the seed URLs. The collaborative crawler is wellsuitable for searching those broad-topic scholarly Web resources.The dissertation makes a comprehensive research on the application of automatic focused searching with the following research methods: literature survey, decomposition & composition, experimentation, et al. According to my research, the paper can provide technical and methodological support for automatic construction of Web resource in the Digital Library. It can theoretically and practically be proved that automatic Web resource construction is feasible and effective.This dissertation has 60 figures, 10 tables.