Semantic Relativity Calculation Based On Concept Lattice And Its Application

Search engine system, as an Internet information retrieval tool, of which the utilization rate is getting higher and higher. However, in the process of getting accurate information, usually the vast amounts of unrelated results are searched for the users, since the traditional search engine systems are all based on keyword matching search technology. To enhance users'satisfaction towards the Retrieval System, the expansion of semantic query and the support of dynamic query is to be achieved in the system.Semantic relativity reflects the interconnectedness degrees of two words. The related research has been widely applied in the field of natural language processing. The applications of semantic relativity calculation have an important practical value in many fields. Such as: information retrieval query expansion, the elimination of ambiguity, improving information retrieval precision and recall rates etc. Its usage in the research of IR systems can raise the semantic information processing capability to the query, thereby it will improve the efficiency of retrieval, and make the system more intelligent.First, this paper describes current situations of search engines, analyzes the features and shortcomings of existing search engine systems, and presents improving methods. Introduced Intelligent Search Technology, we know that the design and implementation of query expansion in the search engine system is necessary.Second,this paper introduces semantic relativity calculation methods as well as the advantages and disadvantages of them. At present, the relativity calculation is based on the methods of word similarity calculation. Semantic similarity calculation methods can be roughly divided into two categories: a world knowledge-based method, the other is the method based on large-scale Corpus.Last, a semantic relativity calculation method based on concept lattice is proposed. This paper describes the implementation process and design of the method in the FCA search engine system. This method breaks through the traditional design ideas, for the usage of information clustering of concept lattice Characteristics. And it also broadens the application of concept lattices.The contributions of this paper are as follows:1. Researching semantic relativity calculation of Chinese words and related algorithms, analyzing the implementation and application of the relativity calculation.2. Proposing a semantic relativity calculation method based on concept lattice, in which knowledge clustering structure is used. The method is proved to be practicable through the experiments.3. Applying in the FCA search engine system, the method provides query expansion function for the system.