Concept Lattice-based Data Mining Research And Realization

Along with widespread application in Internet, leading to large database became more and more, in order to obtain valuable information for the owner of database from the data, data mining arise at the historic moment. Because data mining gains more application, at present it has exciting study prospect.In 1993, Rakech Agrawal at IBM Almaden Research Center makes association rules mining. Association rules is one of many function at data mining, is more important, more active, more widespread application way, so it as for an important pattern at data mining, is doubly paid close attention.Concept lattice is a mathematics implement founded with conception from data; it can show an arrangement of ideas relation between conceptions which is passed by Hasse. As for an implement which is putted into use at data analyze, it is used widely at information retrieval, at figure library, software engineering and knowledge finding ect and is taken seriously by scholar of home and abroad.This paper' focus is on data mining based on concept lattice and the key is the Association Rules based on concept lattice RRECL (RRECL lattice) of the List Index, the context was involved with generation of concept lattice based on List Index and on this basis of the Mining Association Rules.First of the paper discusses the Research and the basic tasks of data mining, The focuses of discussion are the association rules algorithm a, the realization and optimization of the classical algorithm Apriori, and the need to pay attention in Association Rules and application fields .On these basis, The paper analysis the relationship between RRECL lattice based on List Index and association rules, and describes the theoretical framework of the association rules that based on RRECL lattice of the List Index. The paper discusses idea that generates frequent concept lattice by the model of the RRECL lattice based on List Index. Construction algorithm given, and realized the Association Rules based on RRECL lattice of the List Index, Comparison with Apriori that is the classical association rules algorithm, By the experiments , proved the algorithm's the correctness and superiority.