Optimizing The Classification System Based On Rough Set Theory

Every day, we are observing that a wide-ranging amount of data and information is being stored every moment. Mostly the information is dynamic and transactional, and each time the data is being updated from time to time. Getting knowledge from this kind of huge and dynamic data is really becoming tough task with respect to time and efficiency. Fortunately we are served with the intelligent procedures and methods like neural networks, expert systems, fuzzy sets. Along with them we have the rough set theory is a makeable improvement technique was used for knowledge discovery for data dependencies, associations, data reductions and approximation of classification of data from the large data bases. In this paper we would like to propose a dynamic reduction algorithm that can be applicable for producing the reduced attribute set and abridged or condensed reduct set. We would like to analyze and optimize an algorithm which used the abridged reduct set and construct a new classification system. We have considered attribute dependency and attribute significance of Rough Set Theory and are merged for the generation of cohesive reduct set which uses the Particle Swarm Optimization (PSO) algorithm that reduces the complexity of the existing algorithm and attains good accuracy of the Experimental system results. Finally we dealt with a problem of solving the conflicts in decision rules and classification mechanisms with a new case study to different classes. Optimizing The Classification System Based On Rough Set Theory Paper ID IJIFR/ V2/ E8/ 035 Page No. 2644-2649 Subject Area Computer Science & Engineering