Research On Distributed Treatment Of Concept Lattices And Knowledge Discovery Based On Its Framework

Concept lattice also known as Galois lattice, the core data structure of formal concept analysis, is a effective tool of data analysis and extraction rules which uses a binary relation to establish a hierarchical structure of concepts . With the development of the research of concept lattice, it has been used widely in knowledge discovery in database, information retrieval, and software engineering and become a notable research topic in computer science domain. Based on the background, this paper carried out the following studies:Concept lattice is has been successfully applied in many fields such as knowledge discovery by its mathematical nature, but because of its comprehensiveness, the constructive time of the concept lattice has always been the main obstacle of the application of the concept lattice. This paper discusses in detail the existing concept lattice structure algorithm and presents a new batch of concept lattices_pruning algorithm, based on the concept lattice batch generation algorithm through batch tectonic process analysis.With the increasing of the context, the time complexity and the space complexity of the construction of concept lattice will be dramatically increased accordingly. Exploring the use of new methods and means to the concept of lattice structure has become the main topic of the concept lattice. Being diffirent from the the existing concept lattice construct algorighms, using the strategy of divide and conquer is an effective way to solve this problem. After the discussing in detail of the parallel construct algorithm of concept lattice, this paper presents a new concept lattice parallel construct algorithms based on the the idea of the dividing of the closure system of concept lattice. The corresponding closure system of the concept lattice is divided into a number of independent closure system and proceed to the calculation of the concept.Extracting from the database of knowledge discovery rules is the main content and how to reduce redundancy rules extraction and the search space is a hot research topic in the KDD. This paper presents a new association rule mining algorithm based on the dividing of concept lattice. This algorithm maintains the original unchanged information of the database as the basic and calculates frequent itemsets in each of database.The relationship between the concepts is orderly, reveals the generalization and special relationship of the concepts and reflects the concept of hierarchical structure through the Hasse diagram. However, each of the elements in the set of the objects or attributes is incomparable before prescribing the lexicographical order. Considering the concept in the context as the maximal "Rectangle" , this paper studied the graphics of the concept and used it in the proofs of the correlative theorems.