Design And Implementation Of RCP-based FCA Prototype System XDCKS

FCA (Formal Concept Analysis), also known as the theory of concept lattices, wasproposed by Wille R in1982. As an effective knowledge representation and knowledgediscovery tool, FCA has been successfully used in many fields, such as knowledgeengineering, machine learning, information retrieval, data mining, semantic web,software engineering and so on.Since FCA was proposed, there have been many FCA prototype systems, such asGalicia, Erca, Lattice Miner and so on. However, these systems have some drawbacks.Firstly, Function is poor. Such as file types supported and algorithms species are lessand without FCA applications. Secondly, poor extended performance. If the systemneeds to be extended with functions, the re-development work needs to make largechanges to the source codes, which make the development work very difficult.This paper analyzes the main problems of lack of functionality and poor extendedperformance in the existing FCA prototype systems, propose a new FCA prototypesystem, XDCKS, which realize file parsing, concept lattices building, concept latticesgraphic show, association rule mining, etc. And the system also extend supported filetypes and algorithm species. New system employs Eclipse RCP plug-in technology, andis divided into kernel layer, module layer, and algorithm layer. Modules are divided intoseparate plug-ins to achieve a "plug and play" effect, functional modules can access tothe system by the way of configuration and implementation of extension points. Whenthe system needs re-development, function modules can be added to the system directlyin a certain written form without modifying the original system. It makes this tool hasgood flexibility.