中文摘要：

在安全评价过程中,当构建的指标体系过于庞大,数据繁琐,不利于进行安全评价工作。为了对矿井火灾风险评价指标体系进行快速简化,基于粗糙集理论属性约简功能,采用理论与实例分析相结合的方法,利用Skowron差别矩阵算法对矿井火灾风险评价决策表进行属性约简。研究表明：通过约简,矿井火灾风险评价指标由原来的16个缩减为6个,大大减少了后续评价过程中权重确定的计算量,降低了评价的复杂程度,减少了评估工作量。同时约简结果也为矿井火灾防治工作提供了一些指导意见,体现了差别矩阵在决策表的属性约简中的高效性。

英文摘要：

In the process of safety evaluation,when the index system with complicated data is too big,it is not conducive to safety evaluation work. In order to simplify the risk assessment index system of mine fire rapidly,based on the attribute reduction function of rough set theory,the Skowron discernibility matrix algorithm was used for attribute reduction of decision table for mine fire risk assessment with the combined method of theory and case analysis. The results showed that the risk assessment indexes of mine fire reduced from 16 to 6 through reduction,the amount of calculation for determining the index weight in the process of following assessment was also greatly reduced,and the complexity and workload of assessment reduced after reduction. In the meantime,the reduction results provides a certain guidance for mine fire prevention and control work,and it reflects the high efficiency of discernibility matrix in attribute reduction of decision table.