粗糙集属性约简的基因算法

Reduction of Rough Set Attributes Using Genetic Algorithm

粗糙集是一种新型的数据分析方法 ,而属性约简是其主要算法 ,由于属性约简为 NP问题 ,使得传统算法很难将粗糙集应用于实际中。利用基因算法的搜索寻优技术 ,提出一种基于多目标优化基因算法的属性约简算法 ,它不仅有较高的约简效率 ,而且可以得到多种约简属性集。以柴油机油路故障为例 ,进行特征属性约简 ,取得了较好的结果。

Rough set is a new type of data processing method, and the reduction of attributes is its main tool, which is used to find the simplified combinations of attributes under the conditions satisfying the classification precision. Since the attribute reduction is an N-P problem, traditional method is difficult to be applied in the practical application of rough sets. In this paper, using the optimal search technique, a method of attribute reduction based on the multi-object optimization genetic algorithm is put forward. It improves reduction efficiency, and gets more than one kind of attribute reduction aggregate. Finally, the attribute reduction is applied in the feature selection of oil route fault signal of diesel engine, and a good result is obtained.