| At present, Web with a wealth of information resources has become an important channel for people to access information. However, it is very difficult for users that the information can be searched out from the mass information quickly and efficiently, which mainly dues to the disorder of Web information organization and the defect of search keyword matching. Traditional information search techniques base mainly on keyword matching, it is only on literal matching,and do not have the ability of semantic inference and the function of semantic guide on the user's search request. So it is easy to produce a large number of irrelevant links when users search the information, and users need to spend a lot of time to exclude those unrelated information that can not meet the user's needs, which leads to the recall and precision rates are very low. So how to improve the search quality of Web information becomes an important study topic in the field of the current information search and knowledge management.In order to improve the quality of Web information, it is necessary to order the information on the Web, and let the computer understand the information on the Web, and effectively regulate the user's search request and finally achieve semantic retrieval. Because on the semantic Web, information is given well-defined, the machine can understand it, so the semantic Web provides a technical basis for Ontology-based semantic retrieval achieve implementation. In addition, the ontology and natural language understanding technologies also provide technical basis for Ontology-based semantic retrieval achieve implementation.Ontology originated in philosophy, is " Systematically describe the objective existence of the world, namely, Ontology ", it is to study the basic characteristics of the objective world of science. Since 1990s, the concept of ontology has been inroduced in the field of artificial intelligence, knowledge management and library information. The concept of ontology according to the different, areas of the corresponding changes occur, there is no unified concept. However, the most widely used is "Is a clear formal specification of a shared conceptual model." This definition includes at least four meanings, conceptualizatoin explicit, formal, shared. Because ontology has a good Concept hierarchy structure and Supports the logical inference, it also can through the relationship between the concepts expressed the concept of semantics, so ontology provides a technical basis for the semantic retrieval.This article first introduces the theory of ontology and methodology, and briefly introduces the basic concepts of ontology, type, Modeling primitives, Construction principles, construction methods and tools. Then introduces the Semantic Web And its relationship with Information Retrieval. Finally, on information retrieval systems theory, common information retrieval techniques and traditional information retrieval systems. Bedding in a series of theoretical basis of the foundation, the article analyzes several problems to achieve ontology-based retrieval. Then study the logical model of ontology-based retrieval system, detailed analysis and design of ontology-based retrieval system sub-module, proposed ontology-based retrieval system logic model diagram.Designed the logical model of retrieval system, it is necessary to validate this model. Therefore, we must first build a domain ontology. This paper studies the theory of formal concept analysis, then proposed the FCA-based ontology, Formal concept analysis is to understand the mathematical concept of philosophy. It from extension and intension to formal description the concept. It use the binary relation to express the background, and can find out all the implied relationships between concepts and concept, and form The hierarchical structure of concept to realiaze the semantics of information that a computer can understand. FCA-based ontology can reduce the developer's subjectivity, and use mathematical symbols to represent those concepts.Application this method constructed a simple animal ontology, then use this ontology to Pretreatment and Indexing the network of "orangutan" and finally retrieved to verify the feasibility of this model.With the bulk of technical and formal concept analysis technology, The writer believe that ontology-based semantic search will be the main methods of retrieval. Limited due to time and knowledge, this animal in the field of ontology construction is relatively simple, the modeling process also has many deficiencies, these will follow the Yanjiugongzuo Zai further improved in. |
