Unification of Knowledge Discovery and Data Mining using Rough Sets Approach in A Real-World Application

Abstract

In today’s fast paced computerized world, many business organizations are overwhelmed with the huge amount of fast growing information. It is becoming difficult for traditional database systems to manage the data effectively. Knowledge Discovery in Databases (KDD) and Data Mining became popular in the 1980s as solutions for this kind of data overload problem. In the past ten years, Rough Sets theory has been found to be a good mathematical approach for simplifying both the KDD and Data Mining processes. In this paper, KDD and Data Mining will be examined from a Rough Sets perspective. Based on the Rough Sets research on KDD that has been done at the University of Regina, we will describe the attribute-oriented approach to KDD. We will then describe the linkage between KDD and Rough Sets techniques and propose to unify KDD and Data Mining within a Rough Sets framework for better overall research achievement. In the real world, the dirty data problem is a critical issue exists on many organizations. In this paper, we will describe in detail how this KDD with Rough Sets approach framework will be applied to solve a real world dirty data problem.