基于粗糙集萤火虫智能优化算法的特征选择  

Feature Selection Based on Rough Set Firefly Intelligent Optimization Algorithm

作　　者：张帆 ZHANG Fan(School of Information Engineering,Gansu Minzu Normal University,Hezuo Gansu 747000)

机构地区：[1]甘肃民族师范学院信息工程学院,甘肃合作747000

出　　处：《甘肃高师学报》2025年第1期11-16,共6页Journal of Gansu Normal Colleges

摘　　要：针对萤火虫算法中步长参数由经验确定导致算法容易陷入无法收敛或最终解不是全局最优解等问题,文章提出基于粗糙集的自适应步长萤火虫算法的特征选择.萤火虫算法可以根据每个目标特征的吸引度和亮度来重新确定信息位置,极大地提高了算法的收敛速度和搜索能力.在UCI数据集进行测试,得到决策树(DT)和支持向量机(SVM)的分类准确率和收敛速度的测试结果,相较于原始萤火虫算法及其他算法,基于粗糙集的萤火虫算法有更高的准确度和更好的收敛度.In response to the problem that the step size parameter in firefly algorithm is determined by experience,which makes the algorithm prone to convergence failure or the final solution is not the global optimal solution,this article proposes a feature selection based on rough set adaptive step size firefly algorithm.Firefly algorithm can redetermine the information position based on the attractiveness and brightness of each target feature,greatly improving the convergence speed and search ability of the algorithm.Finally,the classification accuracy and convergence speed of decision tree(DT)and support vector machine(SVM)were tested on the UCI dataset.Compared with the original firefly algorithm and other algorithms,the rough set based firefly algorithm has higher accuracy and better convergence.

关 键 词：粗糙集 萤火虫算法 特征选择 

分 类 号：TB181[一般工业技术]