论文部分内容阅读
为对煤矿冲击地压危险性等级进行预测,综合考虑煤层厚度、煤层倾角、开采深度、顶板岩性、构造情况、开采方法、有无煤柱、采煤工艺等影响因素.采用局部加权学习方法建立冲击地压危险性等级预测模型,其中分类器选择随机森林,样本间距离采用欧氏距离函数进行计算.实验选取17组冲击地压数据进行研究,其中14组数据用于建立预测模型,采用十折交叉验证法对模型进行评价,并与采用决策树和朴素贝叶斯生成的模型进行比较,预测准确率得到较大提高,最后使用该模型对其它3组测试数据进行预测,预测结果与实际类别吻合.研究结果表明:采用局部加权随机森林方法可以建立泛化性能更好的冲击地压预测模型.
In order to predict the rank of coal mine rock burst pressure, considering factors such as coal seam thickness, coal seam dip angle, mining depth, roof lithology, structure, mining method, presence or absence of coal pillar and coal mining technique, local weighted learning method The prediction model of rockburst risk level is established, in which the classifier chooses random forest and the distance between samples is calculated by the Euclidean distance function.In this experiment, 17 groups of rock burst data were selected for study, of which 14 groups were used to establish the prediction model, Ten fold cross-validation method was used to evaluate the model. Compared with the model generated by the decision tree and Naive Bayes, the prediction accuracy was greatly improved. Finally, the model was used to predict the other three sets of test data. The results show that using the locally weighted stochastic forest method can establish the prediction model of rock burst with better generalization performance.