An Algorithm Research On Data Mining Based On Rough Set

Knowledge discovery is to abstract the previously unknown and potentially useful models from a large amount of data, and data mining is an important component in the procedure of knowledge discovery. It has been becoming hotspots in many fields how to obtain not only the superficial information but also the underneath one from a large scale of data.As a new popular theory in the AI field, rough set can effectively express and reduce incomplete and uncertain knowledge. This feature makes it very useful in the applications of knowledge discovery and data mining.Therefore, based on the theory of rough set, the thesis proposes a new reducting method, which is a key phrase in data mining, discusses inner meanings of contents of discernibility matrixes, analyses the relationships between attribute reducts and value reducts, and designs an optimal reducing algorithm based on discernibility matrixes, which makes the attribute reducing in according with attribute value reducing and the procedure of knowledge reduction relatively simpler.This thesis introduces the design and development of the CIMS engineering financial audit system for the Dalian Steel Group, which consists of primary design, detail design and software development. The system mainly includes the following sub-systems: sales audit, material audit and profit audit. It is running properly in the company.This thesis adopts, based on the discernibility matrixes, the optimal reducing algorithm into the development of the profit audit subsystem. This algorithm obtains a set of rules obtained, which play conductive roles in company sales decision-making.