Study Of Ontology-based Web Information Retrieval System And Related Technologies

In response to the more and more serious situation of "information overload"and "information confusion"on the Internet, intelligent information retrieval and personalized information service currently become the focus of information service studies. Designing rational intelligent information retrieval system, representing and organizing information based on semantic level, and constructing personalized customer interest model are the bases for realizing intelligent information retrieval and personalized information service. Based on reviewing the current status of study in these aspects, this thesis presents an intelligent Web information retrieval system based on ontology and multiple agents, gives its architecture, principle functions and work flow. In this system, information on the Internet are classified and indexed by using of ontology technology and intelligent agent technology. User queries are also normalized so as to achieve the purpose of finding the required information quickly and accurately. In addition, the system integrates various intelligent agents such as task identification agent, information collecting agent, information processing agent, etc. Therefore, the system has many new characteristics and merits. A semantic indexing method based on domain ontology is introduced. That is, mapping relationships between terms and concepts in ontology are established through semantic analysis, concept category of document or fragment of document and its semantic relationship with other category are established through automatic term analysis, and semantic indexing is conducted by using these concepts and concept categories. When user query is being processed, employment of synonym list of concepts can increase recall of information retrieval, and elimination of less relevant and polygamous concepts through context analysis can ensure certain level of precision of information retrieval. By constructing user interest models based on their searching habits automatically recorded. Thus searching accuracy can be increased and personalized searching can be realized.