新型灰狼算法的粗糙集属性约简及应用

Rough set attribute reduction of novel grey wolf algorithm and its application

粗糙集理论是模式识别和机器学习的重要内容,属性约简是粗糙集理论中核心步骤。然而传统的粗糙集理论对数据集进行属性约简,计算复杂度高,容易陷入局部最优解。提出了一种新型灰狼优化算法的粗糙集属性约简技术,可以很好地解决传统粗糙集理论出现的弊端。同时为了验证算法的可行性,采用国际通用UCI数据库进行验证,与两种传统的属性约简方法进行对比分析。实验结果表明,该方法属性约简个数少,识别精度高,证明该方法切实可行,操作简单。

Rough set theory has become an important part of pattern recognition and machine learning,attribute reduction is a core step in rough set theory.However,traditional rough set theory has a high computational complexity and is easy to fall into local optimal solution.In this paper,rough set attribute reduction algorithm of a novel grey wolf algorithm is proposed,which can solve the drawbacks of traditional rough set theory.In order to verify the feasibility of the algorithm,the paper uses the international general UCI database to verify.Then compares it with two traditional attribute reduction methods.The experimental results show that the proposed method has the advantages of less number of attribute reduction and high recognition accuracy.And it is feasible and easy to operate.