A Research On Mining Of Association Rules Based On Concept Lattice And Rough Sets

However, such a large amount of information, obtain useful knowledge has become increasingly difficult. So the effective use and handling large amounts of data play an important role in current worldwide issues of common concern, knowledge Discovery in Databases (KDD), and its core technology - Data Mining (DM) emerge as the times require. Association Rules (Association Rules) is an important area of data mining research branch.A application of Mining Association knowledge is traditional market analysis, analyze customer's shopping habit by discovering the relation being the purchase of goods, and help retailers develop marketing strategies to promote the sale of goods.Currently more and more industrial, commercial, corporate, financial, government departments, mining valuable association rules from its own database to help make reasonable decisions.Concept lattices and rough set are effective methods of data analysis, which has been widely used in machine learning, artificial intelligence, knowledge discovery, and other fields. Rough set makes use of equivalence relation to carry out classification on the data table, but the concept lattice is based on this data table, combines order theory, especially complete lattice theory, to deal with the concept of a layered discussion. The relationship between them has aroused the concern of many researchers. Although data mining research carried out quickly in recent years in China, because of a late start in the great extent, still lag behind foreign countries, especially based on rough sets and concept lattices theoretical study.As people engaged in the study were relatively small in China, there are still gaps in many areas. Therefore, this study can effectively narrow the gap between domestic and foreign, to fill gaps in some places, for the modernization of the decision-making enterprises to provide intelligence on the protection and support.This paper has done the main work can be summarized as follows:1. Apriori algorithm optimization: improving the algorithms, reduce the necessary scanning data set size, bring forward a simple data structure---tree to storage data sets.2. Combine Rough set theory, derived from "multi-attribute indiscernibility relation."3. Mining association rule based on pruning concept lattices.4. Have discussed the relation between rough set and concept lattices,and put forward rough concept lattices.