Document Retrieval System Based On The Model Of The Conceptual Matching Degree

In the traditional manual document retrieval, the indexing and retrieval of scientific documents are done manually. With increase in the number of electronic documents, manual document retrieval has many limits, we must use computers to process this large number of documents so that we can increase the number and the speed of document-processing. Information Retrieval System has been made a large development and progress for decades, Information Retrieval System basing on from easy-to-use Boolean Search Model, the Vector Space Model to the Probability Model, which increases the accuracy of Information Retrieval. But different Information Retrieval Systems have different performances for different documents. It is very useful and necessary to make existing retrieval systems adapt to the Chinese scientific documents in order to improve the accuracy of Information Retrieval.How to improve page representation to make the actual content of the page more accurate and how to improve the matching method between document and user's query, are related to the precise rate of document searching. This paper analyze the document representation from the Concept Lattice theory, through mapping the query and returning documents to the ontology, we can structure the Concept Lattice from their common background, then we use the Conceptual Matching Degree calculating the degree between the concept of the question and any other the concept. Finally, we sort the degree in accordance with the matching degree, so that the answer can meet the user's query. This is a Document Retrieval System based on the model of Conceptual Matching Degree in Chapter 3.Taking into account a document used as an object, a keyword as an attribute of the concept, if we establish a single large-scale Concept Lattice for the whole set of documents, there will exist complex algorithms and large moving expenses among lattices. Therefore, this paper also makes a Document Retrieval System based on the model of Concept Lattice Matching Degree. Examples show that the improved model matches more with the thinking of human's thought and with people's understanding. This model is feasible.