关键词: 数据挖掘, 底板破坏深度, Weka平台, 贝叶斯分类器, 支持向量机, 神经网络, 决策树, 随机森林

Abstract: The research on the depth of destroyed floor in coal mining is important to the realization of coal mine safety and high efficiency mining. In view of the large error problems of single factor predicting the depth of destroyed floor, based on the open source data mining tool of Weka Platform and analyzing of the sample factors, and the data of destroyed floor depth was analyzed by using Bias Classifier, Support Vector Machine, Neural Network, Decision Tree and Random Forest Model. The comprehensive prediction of destroyed floor depth was completed from the perspective of multiple factors. The application results show that the length of working face and mining depth were the main influencing factors on destroyed floor depth; The node error rate of the Neural Network Model was the lowest, and the node error rate of the Decision Tree Model is the best; The accuracy rate of Neural Networks and Random Forests is 95% in detail; With overall analysis, the forecasting effect of Neural Network Model was optimal, and it achieved better prediction of destroyed floor depth.