Scientific Papers Based On The Concept Of Association And Of The Associated Retrieval Research

After briefly describing the basic structure and the classification of search engines, this paper analyzes the development and basic characteristics of popular Chinese and English search engines on Internet at present, summarizes the defects and introduces the research hot spot and foreground, of existing search engines in summary.In the following chapter, after introducing the basic theory of the conceptual network(CN), the paper discusses it's features. Aiming at solving the indeterminacy of knowledge expression, we propose a method to express domain knowledge by means of CN and bring forward strategy of constructing a professional domain conceptual network used to retrieve scientific papers. At the end of this part, the formation and derivation of nodes in the network are studied.On this base, we present two methods for retrieving scientific papers distinguishing^ based on conceptual roles and based on author relevance. We wish we could exalt the efficiency of searching and solve the problems of low recall and low precision when using existing general search engines to search papers by means of the combination of the two methods and CN.On the one hand, the title of a scientific paper is the condensation of its content. Usually, we can get some useful information from the titles such as the subject, the purpose, the technical way, the range and related domains. It is helpful to identify the schemas of papers' titles correctly for determining the categorization of papers. In chapter 3, we discuss how to dig out the in-formation from papers' titles and how to use them to classify papers. On the other hand, the relationship among authors reflects the internal relevance among research personals and among different domains. In chapter 4, we lay emphasis on how to discover these relations and construct a author relevance network(ARN) covering various domains.At the realization of the system, we analyze the integral structure and working principle of our system at first. Then, we show the relationship among tables in core database. Lastly, we study automatic document categorization algorithm and propose algorithm descriptions and experiment results of Chinese language segmentation, schema matching of paper titles andclustering.At last, a conclusion is drawn and a prospect for continuing work of the paper is offered.