Rough set-based approach to data mining

To improve competitiveness, enterprises—big or small—have been trying to use information technology to help in all aspects of their business. The growing volume of data in digital form and advances in data analysis methodologies and information technology have led to a field that attempts to extract useful information and intelligence from these large data sets for the purpose of strategic and decision making, such a rapidly growing field is called Data Mining.; This dissertation focuses on feature extraction and rule induction aspects of data mining based on the so-called Rough Set theory. The first aim of this dissertation is to develop an optimisation feature extraction model, which can reduce the unnecessary input variables by selecting the desired and representative reducts from each object. The second main purpose is to develop a rule-generation algorithm to generate a minimal set of rule-reduct, from which useful knowledge can be induced. Finally, we want to modify the rule-generation to apply to incomplete information systems. The organization of this dissertation is as follows. Chapter 1 discuss the motivation and dissertation outline. Chapter 2 introduces basic theories and applications of Rough sets, feature extraction model and rule-generation algorithm of literature review. Chapter 3 presents the integer programming model for feature extraction problems. Chapter 4 proposes rule-generation algorithm and rule induction for complete information systems. Chapter 5 proposes a generalized rule-generation algorithm and rule induction for incomplete information systems. Chapter 6 presents numerical results. We demonstrate the test results of using examples found in the literature as well as newly constructed. Chapter 7 presents the conclusion and future work that we plan to carry out.