基于粗集和神经网络的煤矿井下危险度评估

Risk Assessment in Coal Mines Based on Data Mining and Neural Networks

结合煤矿探测机器人,提出了一种井下气体危险度评估的新算法:首先利用粗集理论对煤矿井下环境的各类数据进行属性和对象约简,然后把简化后的样本数据输入到神经网络进行训练,利用其输出结果进行系统危险度预测和评价。该算法综合了粗集的简化功能与神经网络分类的强鲁棒性的优点,实验证明其模型结构简单,可有效地评估井下危险度,为环境安全评估建模提供了一种新的途径。

Combining coal exploration robots,a new algorithm of gas risk assessment in coal mines is proposed,first,the at-tribute and object simplification of various data is carried out based on data mining,and then the simplified data are input into the neural net-work to predict and assess the risks.The algorithm combines the advantages of the simplification of data mining and the robustness of neural networks.The result shows that the model,which is simple in structure and can effectively assess the risks in coal mines,can be a new way of environmental safety assessment.