| Along with the high speed development of computer technology and communication technology all over the world, various industries accumulated the massive data. The implicit message of the data is the wealth of enterprises. Data mining can get the connection and the implicit information of the data, predict the future trend of development.Rough set is an important data mining method. Its characteristic is not to need any priori information. Only according to the data itself, through the data reduction, we can draw valuable decision rule.Attribute reduction is one of the important links of rough set theory. Attribute reduction can keep the information system classification ability unchanged, delete the redundant attributes information, improving the information system of decision making efficiency. This paper based on the characteristics of several common attribute reduction algorithm, proposed an improved reduction algorithm based on dichotomy and weighed sum of attribute significance.Finally use Changchun City Administration for Industry and Commerce's registration data of all kinds of market entities as the main research object. Use the information of Registered, cancelled or revoked and the annual check-up as the carrier of the data mining.Establish the SMEs enterprises information analysis data mining model. Combined with the proposed improved reduction algorithm, we get the decision rules of the risk warning for SMEs, and verify the feasibility and effectiveness of the algorithm. |
