论文部分内容阅读
为更准确地预测瓦斯涌出量,预防瓦斯灾害,有必要建立和应用基于改进极端学习机(IELM)的混沌时间序列预测模型。首先,对瓦斯涌出量监测数据构成的多变量时间序列进行相空间重构,采用互信息法与虚假邻点法得到每一变量的延迟时间和最佳嵌入维数;然后,通过最小二乘方法和误差反馈原理计算出最优的网络输入层到隐含层的学习参数,对极端学习机(ELM)进行改进;最后,借助IELM建立瓦斯混沌时间序列的预测模型。通过仿真试验,运用该预测模型预测的最大相对误差为3.290 2%,最小相对误差为0.898 2%,平均相对误差为1.952 8%。
In order to predict gas emission more accurately and prevent gas disasters, it is necessary to establish and apply a chaotic time series prediction model based on Improved Extreme Learning Machine (IELM). Firstly, phase space reconstruction is performed on multivariable time series composed of gas emission monitoring data, and the delay time and optimal embedding dimension of each variable are obtained by mutual information method and pseudo-neighbors method. Then, Method and error feedback principle to calculate the optimal network input layer to the hidden layer of the learning parameters, to improve the Extreme Learning Machine (ELM); Finally, with IELM to establish gas chaotic time series prediction model. The simulation results show that the maximum relative error predicted by this prediction model is 3.290 2%, the minimum relative error is 0.898 2% and the average relative error is 1.952 8%.