The Research Of Knowledge Discovery Based On Domain Knowledge And Concept Lattice Model

Knowledge discovery in databases (KDD) is a focus topic in research fields of databases and Artificial Intelligence, which aims to nontrivial of mining the interesting potentially useful, valid and understandable knowledge from data. The high-efficient approaches are the research focus. However, the most of systems of data mining start the mining from scratch. In another word, the process of mining is independent of the domain knowledge of certain field. More research has shown that the domain knowledge is important for data mining.Concept Lattice represents knowledge with the relation between the intensions and the extensions of concepts, and the relation between the generalization and the specialization between concepts, thus it is an efficient tool for KDD. By Introducing equivalent intension into GCL, the Extending model of Concept Lattice is gotten which represent the knowledge more clearly and distinctly. It is widely used in information retrieval, software engineering and KDD. However, the size of ECL is too large owing to its completeness, which make it unsuitable for large database.The domain knowledge is introduced into data mining in the dissertation, and the role of domain knowledge in data mining and its prospect is also discussed. Integrating and Updating domain knowledge with the process of data mining, and merging knowledge into the original knowledge base, will contribute to reducing the scale of the data and improve efficiency and quality of knowledge that will be mined. In the dissertation, several algorithms are presented for the decreasing of ECL, which use domain knowledge to induct the values of each attribute, and climb to its exact level. Algorithm DKLT and DKLTM are based on the domain knowledge in the form of concept hierarchy tree, and algorithm DKLG is based on concept hierarchy graph. These approaches can compress the data, increase the efficiency and quality of discovery knowledge, and it is helpful to solve the certain problems of KDD in the large databases.A next-generation data mining environment should actively support a user to both incorporate the domain knowledge into mining process and update this domain knowledge with mining results. The domain knowledge will play a role in the data mining environment as important as the data mining methods.