中文摘要：

煤矿监测数据实质是一种数据流，煤矿安全评价可以看作是数据流的分类，分类的标识为安全和不安全。在随机决策树模型的基础上，使用 Hoeffding Bounds 不等式与信息熵确定分割点，代替用随机选择方法确定分割点。实验结果表明该方法对数据流分类具有更好的分类精度，为煤矿安全评价提供了一种新的实用方法。

英文摘要：

Monitoring data in coal mine is essentially a data stream. Coal mine safety evaluation can be seen as the classifica-tion of the data stream, and classification categories are safety and unsafety. Based on random decision tree, a method was proposed in the paper, and the method determined the split point by Hoeffding Bounds inequality and information entropy instead of random selection. Experimental results showed that the method has better accuracy for data stream classification. Therefore, a new practical approach is provided for coal mine safety evaluation.