Research On The Key Technology Of Concept Lattice In Data Mining

The computer technology has greatly developed and been applied in daily life, which results in the increase of huge databases. The data mining has emerged in order to excavate the valuable information in the database .Data mining has gained increasingly studying in recent years and has been successfully used in widely field.Concept Lattice is a powerful data analysis tool, and used for many tasks in machine learning. At present, the Concept Lattice has been applied in the information retrieval, digital libraries, software engineering, data mining and knowledge discovery, and other aspects, which obtain great attention of scholars.The data mining method based on the Concept Lattice is constituted by the two key steps: construction of concept lattices and extraction of designed grid knowledge. First, construct the concept lattice to reflect the connections, that is contained and being contained, between concepts of different levels contained in the data. Then excavate the associating knowledge, classification knowledge and clustering knowledge in the concept based on the cases and the special relationship between thethe concepts.This paper is mainly concerned about the research of Concept Lattice in data mining. we introduce the related theory of Concept Lattice firstly and then mainly discusse the relationship between Concept Lattice and rough set, which has laid the theoretical foundation for understanding of rough set and Concept Lattice and still make preparations for better using of rough set and Concept Lattice to analyse the data. Moreover, we discusse the extraction of association rules in data mining and give the improved ApriorTid algorithm, which is improved effective. What's more, we bing forward the structure algorithm of the index Concept Lattice based on improved Godin and corresponding algorithm of rule-excavating, which combines the improvement of lattice structure and quantitative pruning and greatly improved the efficiency of extraction. Then we bring foward the method which extract decision rules based on pruning quantify Concept Lattice, the structure of Concept Lattice makes full use of the improved algorithm superiority, at the same time we take into account the Concept Lattice generation without worthless nodes.