The Research And Application Of Credit Rating Of Corporate Environmental Behavior Based On Decision Tree

Abstract:With the growth of environmental awareness, people began to pay attention to the environmental behavior of enterprises. However, there still existed some new problems in the daily environmental supervision of the environmental protection administration. So environmental protection department started a rating job on the environmental behavior of enterprises and announced corporate environmental behavior. It will compel enterprises to accept public supervision. At the same time, the results of the rating are naturally associated with the economic means such as financing credit so as to urge enterprises to strengthen pollution treatment, reduce emission of pollution, and obey the law consciously.According to the relevant national laws, regulations and the actual situation of Hunan province, the thesis put forward a credit evaluation framework on corporate environmental behavior based on decision tree. Firstly, it sums up the rating index system of enterprise environmental behavior, what’s more, it uses roughs set theory to reduce the index system and get the reduction index system at last. This thesis presents two meliorations aiming at decision tree fragments and continuous attributes discretization after the analysis of shortages of traditional C4.5algorithm. Experiments show that improved algorithm has a certain degree of improvement in the node number, running time, and the classification accuracy. This thesis provides an optimized C4.5algorithm through combining the improved algorithm with rough set theory. The algorithm uses rough set theory to reduce the attributes in the decision system, and uses the reduced data as the input of C4.5algorithm to construct a decision tree. This article has put this new algorithm into practice, and the result of the experiment shows that the optimized algorithm has higher efficiency and accuracy compared with the traditional C4.5algorithm. Finally, the thesis makes a data mining to the data of enterprise environmental behavior and gets a satisfied result. The experiment makes it clear that we could apply the credit rating framework of corporate environmental behavior to the credit rating system of corporate environmental behavior.