الفهرس | Only 14 pages are availabe for public view |
Abstract For the past two decades there has been a dramatic increase in the amount of information or data that is collected and stored electronically. With this trend continuing, the problem in present times has become what to do with this valuable resource of data? Information is at the heart of any successful business operation and decision-makers use data stored to gain valuable insight into their business processes. Some of the common techniques in use presently are statistics, visualization, neural networks, decision trees and rule discovery. All these techniques have their individual strengths and weaknesses and so one must combine them appropriately to exploit their strengths to achieve high quality knowledge. One common problem faced by these techniques is noise. Noise is the errors in the information contained within the data. It leads to uncertainty, which can eventually and significantly affect any knowledge that is mined from data. In this thesis, we discuss several techniques that are used for knowledge discovery. We show their individual strengths and weakness. We propose two approaches for knowledge discovery for complete information systems based on rough sets theory. Because well preprocessed data is essential for important rule generation, we propose a new approach for knowledge discovery for incomplete information systems. These approaches are used in a wide variety of occupations, it helps people to identify, study, and solve many complex problems which enable the decision makers to make informed and better decisions about the uncertain situations. |