煤层顶板冒落预测方法分析

Analysis of methods to predict the seam roof caving

本文以煤层顶板冒落数据作为样本,利用weka等软件,利用粗糙集理论等相关非线性理论知识,分别采用支持向量机和Bp-神经网络对煤层顶板冒落进行训练和预测。从详细的精度,混淆矩阵和节点错误率这三个方面分别比较三种算法,从而得到结论是:Bp-神经网络的训练和预测效果是稍微强于支持向量机。

In this paper, seam roof caving data as samples using weka software, use of relevant knowledge of the nonlinear theory of rough set theory, were used SVM and neural network Bp-seam roof caving training and prediction. From the detailed precision, confusion matrix and the error rate of these three aspects node compare three algorithms were to obtain the conclusion is: the training and prediction effect Bp-neural network is slightly stronger than the support vector machine.