The Research And Application Of Death Rate Per Million-ton Coal Prediction Method

By analyzing the calculation methods and problems of the death rate per million-ton coal(DRPMTC), the paper adopts the gray relational analysis method and the two improved gray relational analysis methods(the data dimensionless method based on Vague ideas and set [-1,1] linear generate operator and association weighted based on coal production) to establishe the index system of DRPMTC with the data of each coal-producing province in recent years. It contains the proportion of collective ownership coal production, the proportion of high-gas coal mine, the proportion of coal and gas outburst coalmines,the proportion of the engineers and technicians, mechanized tunneling rate, rate of mining coal mechanization, the average wage, mechanized mining coal rate and the full efficiency.Then, the dissertation introduces the gray prediction model based on the buffer operator, and calculates the prediction indexes of DRPMTC, and selects the best estimated values from the prediction values of the previous10order buffer operator. Finally, a new combination prediction model,Dm-GM(1,1)-LSSVM, which combines gray prediction model and LSSVM prediction model is put forward. directly affect the accuracy of support vector machine. The dissertation uses genetic algorithm (GA),particle swarm optimization (PSO) and grid search algorithm for the optimal penalty parameter and nuclear parameter of the prediction model,and improves the accuracy of death rate per million-ton coal with examples.