基于人工神经网络的煤矿安全事故预测研究

A research on coal mine safety forecast model based on artificial neural networks

结合2000~2009年间的全国煤矿统计数据,建立了预测模型,以计算机软件MATLAB 7.5为平台,结合BP神经网络学习算法对10年间的煤炭产量、死亡人数和百万吨死亡率进行了预测分析。结果表明,检测样本的期望输出与模拟输出线性回归分析时的线性相关系数均在0.91以上,说明了BP神经网络较好的拟合效果。

On the basis of the statistic data of coal mines in China from 2000 to 2009,a forecast model is established.With computer software MATLAB 7.5 as the operating platform and BP neural network learning algorithm,the authors present in this paper a forecast of coal production,death toll and mortality rate per million tons of coal produced during this ten-year period.The results show that linear correlation coefficients of the expected output of the samples tested and the simulation output linear regression analysis are all above 0.91.This indicates that BP neural network algorithm（ANN） have a good application prospect over safety forecast in coal mines.