改进蜻蜓算法及其在特征选择中的应用  

Improved dragonfly algorithm and its application in feature selection

作　　者：王万良[1] 朱凯莉 李伟琨[1] 赵燕伟[2] 介婧[3] WANG Wanliang;ZHU Kaili;LI Weikun;ZHAO Yanwei;JIE Jing(College of Computer Science&Technology,Zhejiang University of Technology,Hangzhou 310023,China;College of Mechanical Engineering,Zhejiang University of Technology,Hangzhou 310014,China;School of Automation and Electrical Engineering,Zhejiang University of Science&Technology,Hangzhou 310023,China)

机构地区：[1]浙江工业大学计算机科学与技术学院,浙江杭州310023 [2]浙江工业大学机械工程学院,浙江杭州310014 [3]浙江科技学院自动化与电气工程学院,浙江杭州310023

出　　处：《计算机集成制造系统》2020年第8期2124-2132,共9页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金资助项目(61572438,61873240)。

摘　　要：为合理应用生产制造中产生的数据,分析和挖掘出数据中的关键特征和潜在信息,以帮助企业提高效率、降低成本,提出一种改进的蜻蜓算法,并将其用于特征选择。算法在继承蜻蜓算法良好收敛性的基础上,引入局部序列浮动后向选择机制,在增加算法的全局搜索能力的同时提高了算法的求解精度。在标准复杂数据集上对该算法与其他特征选择算法进行仿真对比与分析,并将其应用在半导体生产线上进行验证。实验结果表明,所提算法无论是在标准数据集上还是实际工程问题上,均表现出良好的全局搜索能力与发展潜力。To properly apply the data generated in the manufacturing process,analyze and mine the key features and potential information in the data to help the enterprise improve efficiency and reduce costs,an improved dragonfly algorithm was proposed for feature selection.On the basis of inheriting the good convergence of dragonfly algorithm,the proposed algorithm introduced a local sequence floating backward selection mechanism,which could increase the global search ability of the algorithm and improve the accuracy of the algorithm.The algorithm was compared and analyzed with other popular feature selection algorithms on some standard complex datasets,and applied to real semiconductor production line for verification.The experimental results showed that the proposed algorithm had good global search ability and development potential in both standard data sets and real engineering problems.

关 键 词：特征选择 蜻蜓算法 序列浮动后向选择 群智能算法 优化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP278[自动化与计算机技术—计算机科学与技术]