Basic Education Domain Ontology-based Intelligent Search Agent Model

The first decade of 21th century has gone and a large amount of valuable web resources about elementary education has sprung up in the Internet with the advance of educational informationization. People who are related with elementary education, including teachers and students in elementary or secondary schools, parents whose children are receiving compulsory education, elementary education administrators working in government departments, are paying more attention on such resources. Accordingly, research works about how to acquire elementary education resources from the Internet effectively have an important significance in increasing teachers' efficiency, broadening students' knowledge outlook and improving the educational informationization process.In order to overcome the weakness of the traditional web search engines in acquiring education informational topic resources, this paper tries to solve the education informational topic information retrieval(IR) problems by taking advantage of the intelligent IR agent technology. An intelligent IR agent model based on elementary education ontology (called OISA-EDU for short) is proposed in this paper. This model refers to those important research results in the IR field adequately, and has advantages such as autonomy and individuation. Users' IR preference which will be used to support a serial of personalized IR services can be captured by the model through a simple interactive interface. The main works of this paper are listed below:(1) Learn the present research situation of home and abroad about intelligent IR agent meticulously, and then an intelligent IR agent model based on elementary education ontology is proposed in this paper. This model is designed to offer an information acquisition channel that is more exact and personalized for people who are related to the elementary education field.(2) Study the ontology based IR technology deeply and construct an elementary education ontology (called EDU-ONT) with protege and OWL-DL on the base of the CELTS-42 and Education Thesaurus. Then this paper proposed a semantic IR model used for elementary education field. Those key technologies affecting the precision and recall are discussed thoroughly, and a lightweight degree semantic annotation arithmetic based on EDU-ONT, uer's query preprocessing arithmetic and method to calculate the result' similarity are introduced successively. From the experiment's results we can conclude that the semantic IR model can improve both the retrieval's precision and recall in elementary education field.(3) In order to satisfy users' personalized information need, this paper designs a personalized IR service system based on the user profile, which consists of three sub functions including personalized information filtering, interested documents recommendation and user customized resource monitoring. This personalized IR system is a supplement for the semantic IR function, which will improve the intelligent agent model's adaptivity for different users.(4) Conclud a useful content organizing rule from elementary education web sites carefully, based which a strategy is proposed that can find the new resources from a web site affectively and exactly.(5) With the help of the open source framework JADE, we finally develop the prototype of OISA-EDU using JAVA; advantages of the intelligent IR agent are shown by the last experiment.