【摘要】 瓦斯涌出量与其影响因素之间存在着高度的非线性性,为了提高瓦斯涌出量的预测精度,提出了变量扩维-筛选法,即引入非线性项。采用混合粒子群算法从大量的候选变量中选出最优的变量子集,利用最优变量子集建立拟线性瓦斯涌出量预测模型。将测量数据分为建模数据和测试数据,测试结果表明基于变量扩维-筛选法的预测模型具有较高的精确度,测试结果优于逐步回归法、神经网络法以及纯粹的线性回归法,具有一定的应用价值。更多还原

【Abstract】 There exist nonlinear height between abstract gas emission and its influencing factors,it provided variable augmented screening method in order to improve the prediction accuracy of gas emission,namely nonlinear term.Hybrid particle swarm optimization algorithm with variable subset from the candidate a number of variables in the optimum,the optimal subset of variables to establish the linear gas emission quantity prediction model.The measured data will be divided into modeling data and test data,test results show that the prediction model of variable dimension variable augmented-screening method based on high precision,linearity test results are better than the stepwise regression method,neural network method and pure regression method,and it has certain application value.更多还原