基于改进的二进制蚁狮算法的特征选择模型及应用  

Feature selection model based on improved binary ant lion optimizer and its application

作　　者：赵转哲[1,2] 叶国文 张宇[1] 刘永明 张振[1,2] 何康 ZHAO Zhuanzhe;YE Guowen;ZHANG Yu;LIU Yongming;ZHANG Zhen;HE Kang(School of Mechanical Engineering, Anhui Polytechnic University, Wuhu 241000, China;Anhui New R&D Institutions of Huamn-machine Interaction and Collaboration, Anhui Polytechnic University, Wuhu 241000, China;School of Mechanical and Electronic Engineering, Suzhou University, Suzhou 234000, China)

机构地区：[1]安徽工程大学机械工程学院,安徽芜湖241000 [2]安徽工程大学人机自然交互和高效协同技术研究中心安徽省新型研发机构,安徽芜湖241000 [3]宿州学院机械与电子工程学院,安徽宿州234000

出　　处：《计算机集成制造系统》2021年第7期1898-1908,共11页Computer Integrated Manufacturing Systems

基　　金：安徽省自然科学基金面上资助项目(1808085ME127);安徽工程大学引进人才科研启动基金资助项目(2019YQQ004);安徽工程大学校级科研资助项目(Xjky019201905);高校优秀青年骨干人才国外访问研修资助项目(gxgwfx2019041)。

摘　　要：为了降低滚动轴承故障数据集的特征维度,选取最有效的数据特征,首先提出一种改进的二进制蚁狮算法,该算法通过引入种群保护集机制,对具有寻优潜力的部分蚂蚁进行保留,并将保护集内群体与主群并行迭代,以加强算法的全局寻优能力,然后通过0-1背包问题验证了该算法的有效性;最后将上述改进融入混合式特征选择模型中,在UCI标准测试数据集与凯斯西储大学滚动轴承故障数据集上分别应用该模型进行特征选择。实验结果表明,融合改进二进制蚁狮算法的混合式特征选择模型的识别精度与特征约简能力均得到明显的提升。To reduce the feature dimension of rolling bearing fault data set and select the most effective features of data,an Improved Binary Ant Lion Optimizer(IBALO)was proposed,which introduced the mechanism of population protection set to keep the ants with optimized potential into protection set.The group in the protection set was iterated parallelly with the main group in order to enhance the global optimization capability.Then the validity of IBALO was verified by 0-1 knapsack problem.The above of improvements were integrated into the hybrid feature selection model,and applied to UCI standard test databases and rolling bearing fault databases from Case Western Reserve University respectively.The experimental results showed that the recognition accuracy and feature reduction ability of the hybrid feature selection model based on the IBALO were improved significantly.

关 键 词：二进制蚁狮算法 种群保护集 并行迭代 0-1背包问题 混合式特征选择 轴承故障 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TH165.3[自动化与计算机技术—控制科学与工程]