Application And Research Of Data Mining In Comprehensive Statistic System For Coal Enterprise

Coal is one of the critical energy for our country. Comprehensive statistics for coal industry is not only one of the foundations for the national comprehensive statistics, but also an important ingredient for the national energy statistics. As the most essential administration work for the coal industry, it provides the basic reference for all decisions, associations and management of each business factor. However, the existing system that currently used has various limitations and shortcomings in terms of functionality and stability, which can not provide deeper data mining and complex decision making. These limitations have directly influenced the quick and scientific decision-making as well as the improvement of administration. Therefore, to research and develop intelligent and comprehensive statistical system for coal industry has both theoretical and practical signification.To tackle the shortcomings of the existing coal comprehensive statistical system, the project establishes a three -layer coal comprehensive statistical system by introducing data mining technique, i.e., adding a data mining layer on top of traditional system. Through thorough comparison between decision tree method, neural network approach, Bayes network method and rough sets, the article puts forward a multiple variant decision tree method combining decision tree and rough set approaches.Prior to the establishment of decision tree, Using rough set theory condition attributes of decision-making table to finish attribute reduction of the property, after the collection, for one thing,to maintain quality of the same classification, and another,the attribute set does not contain redundant attributes.Will be obtained after reduction of condition attributes set for the structure, decision tree, this will help reduce the size of tree to be built. In the establishment of decision tree process, the algorithm for reduction of the number of post-condition attributes, the distance function and the two equivalence relations in the concept of generalized solution of the multi-variable testing problems was used in order to get smaller, higher accuracy decision tree. The algorithm is used to build a coal-effective decision-making model for mining companies to realize the benefits of a comprehensive evaluation of mine. According to the evaluation results, the leaders can see the factors which have an impact on the mine efficiency indicators, so that they promptly take measures accordingly. The establishment of the system provides the rapid data support for enterprise strategy making and production managing, thereby improves the management level of coal enterprises, further upgrades the competition ability of coal enterprises.