摘要
为提高煤矿绝对瓦斯涌出量预测的可行性与准确性,将因子分析法与BP神经网络方法相结合,提出一种改进的BP神经网络预测方法。使用因子分析法对13个煤矿绝对瓦斯涌出量影响因素的原始数据进行降维数据处理,得到3个公共因子;以3个公共因子代替原有13个煤矿绝对瓦斯涌出量影响因素作为BP神经网络的输入层参数,建立因子分析法与BP神经网络法相结合的煤矿绝对瓦斯涌出量预测模型。选取实例数据对改进的BP神经网络预测方法进行验证,最终验证结果:15组训练样本预测值与实际值的相对平均误差为1.39%,证明训练完成的改进BP神经网络模型具有良好的拟合效果;5个预测样本的相对误差均小于2.25%,证明改进的BP神经网络预测模型具有良好的预测准确性。
In order to improve the prediction feasibility and accuracy of absolute gas emission in coal mine,an improved BP neural network prediction method was proposed by combining the factor analysis method with the BP neural network method.Dimensional reduction data processing on the original data of 13 influence factors relating to the absolute gas emission in coal mine was carried out by factor analysis method.3 common factors were obtained,which were applied to replace the 13 influence factors as the input layer parameters of BP neural network.Then,the prediction model of absolute gas emission in coal mine was established by combining factor analysis method with the BP neural network method.The improved BP neural network prediction method was verified by selecting the case data.And the final verification results were obtained as fallows.The relative average error between the predicted value and the actual value of 15 groups of training samples was 1.39%,which proved that the improved BP neural network model had a good fitting effect.And the relative errors of five predicted samples were all less than 2.25%,which proved that the improved BP neural network prediction model had a good predictive accuracy.
作者
马晟翔
李希建
MA Shengxiang;LI Xijian(Mining College, Guizhou University, Guiyang, Guizhou 550025, China;Engineering Center for Safe Mining Technology Under Complex Geologic Condition, Guiyang, Guizhou 550025, China;1nstitute of Gas Disaster Prevention and Coalbed Methane Development of Guizhou University, Guiyang, Guizhou 550025, China)
出处
《矿业研究与开发》
CAS
北大核心
2019年第10期138-142,共5页
Mining Research and Development
基金
国家自然科学基金面上项目(51874107)
贵州省科技计划项目(黔科合平台人才[2018]5781号)
