International Journal of Data Mining & Knowledge Management Process (IJDKP) Vol.2, No.5, September 2012 DOI : 10.5121/ijdkp.2012.2507 83       Ahmed T. Shawky 1 , Hesham A. Hefny 2 and Ashraf H. Abd Elwhab 3 1 Computer Sciences and Information Department, Institute of Statistics and Research, Cairo University, Egypt ah_taisser@yahoo.com 2 Computer Sciences and Information Department, Institute of Statistics and Research, Cairo University, Egypt hehefny@ hotmail.com 3 Computer Sciences and Systems Department, Electronics Research Institute, Cairo, Egypt awahab@ad.gov.eg ABSTRACT Decision models which adopt rough set theory have been used effectively in many real world applications. However, rough decision models suffer the high computational complexity when dealing with datasets of huge size. In this research we propose a new rough decision model that allows making decisions based on modularity mechanism. According to the proposed approach, large-size datasets can be divided into arbitrary moderate-size datasets, then a group of rough decision models can be built as separate decision modules. The overall model decision is computed as the consensus decision of all decision modules through some aggregation technique. This approach provides a flexible and a quick way for extracting decision rules of large size information tables using rough decision models. KEYWORDS Rough sets, Fuzzy sets, modularity, Data mining. 1. INTRODUCTION During the last two decades, rough set theory has received much attention as a promising technique for data mining. Its ability to deal with imperfect data analysis makes it quite efficient for drawing useful decisions from datasets of various real world applications. However, rough decision models suffer the problem of high computational complexity of extracting decision rules when dealing with large-size datasets. Several researchers have proposed variuos approaches to overcome this problem. One of the recent attractive approaches, has been suggested by C. Degang and Z. Suyun [1], to improve the performance of rough decision making process, by integration of fuzzy set and rough set theories. This approach provides two benefits; the first is turning the continuous-valued conditional attributes into nominal or ordinal values which greatly simplifies the computations of finding reducts. The second benefit is the ability to deal with attributes with uncertain values which can be handled with fuzzy linguistic values that differ from decision maker to other. In this paper we make a further step for decreasing the computational cost of finding rough decision rules by introducing the approach of modularity. Modular approach to decision making uses the central idea of task-decomposition to reduce the computational cost of drawing decisions over large datasets. Modular neural networks are successful applications of such an approach [2, 3]. In this paper, we show how this approach can be adopted to reduce the computational cost of large rough decision models. This paper is organized as follows: section 2