摘要
为了有效挖掘影响瓦斯涌出量的各因素之间存在的复杂非线性关系,利用基因表达式编程算法建立瓦斯涌出量预测模型,采用多元回归法得到初始系数和常数,克服了GEP算法随机确定常数的弱点.与GP、BP神经网络、小波神经网络方法相比,GEP模型具有更高的预测精度和稳定性.
In order to find complex and nonlinear relationship among factors which influence gas emission, a method utilizing gene expression programming to set up mine gas emission forecasting model is proposed. The model improves the weakness of GEP in deciding const data randomly, by using multi - regression to find initial coefficient. Compared with GP, BP neural network and wavelet neural network, the forecast model based on GEP has higher precision and stability.
出处
《山东师范大学学报（自然科学版）》
CAS
2015年第2期27-30,共4页
Journal of Shandong Normal University(Natural Science)
基金
山东省高校科技发展计划项目（J13LC51）
山东省科技厅项目（2011XH17006）
山东大学高校院所自主创新计划项目（201401213）.
关键词
基因表达式编程
瓦斯涌出量
预测
gene expression programming
gas emission
forecast
