基于粗糙集的模糊模型辨识方法

提出了一种基于粗糙数据分析的模糊模型辨识方法·二值化数据过滤和决策表的简约算法是该方法的两个关键点·通过将传统的决策表转化成二进制决策表,并采用二值化数据过滤技术,可以同时简化决策表的属性和属性值·通过决策表的简约算法,从决策表中提取出重要的属性和关键的属性值,从而得到了输入空间的模糊最优化分,进一步得到模糊模型的前提结构和参数·利用这种方法以一组经典数据为背景建立了岩石边坡工程中边坡稳定性分析的模糊模型,仿真结果说明了该方法的有效性

Identification of Fuzzy Model Based on Rough Sets Theory

An identification method for fuzzy model based on rough sets data analysis was proposed. There are two key points in the method. First,the traditional decision table is converted to binary decision table, so the attributes and attribute values in decision table can be dealt at the same time by binary data filter technique. Secondly,the significant attributes and the key attribute values, which are related to the optimized fuzzy partitions of input space,are extracted by using reduction algorithm. So the premiss structures and parameters of fuzzy model are got. The method was illustrated using real data from rock slope practice. The fuzzy model for rock slope stability analysis was built,and the simulation results are reasonable.