计算机应用研究2017，Vol.34Issue(12)：3576-3579,3602,5.DOI:10.3969/j.issn.1001-3695.2017.12.013
基于自适应步长的萤火虫划分聚类算法
Firefly partition clustering algorithm based on self-adaptive step
摘要
Abstract
In many areas,clustering is one of the most important techniques,including data mining,pattern recognition and image analysis.Due to K-means algorithm is easy to fall into the local optimum by the selection of initial clustering center,this paper proposed an improved algorithm based on the combination of firefly algorithm and K-means algorithm,which was called ASFA.By using random and global search of firefly algorithm,it initialized the original cluster centers,which could be further used to obtain more accurate clustering of K-means.In the clustering optimization algorithm,it utilized adaptive step size instead of the original fixed step size to avoid local optimization of the algorithm and obtained higher accuracy.In order to improve performance,it implemented the new algorithm in benchmark datasets of different size.The experimental results show that ASFA has better clustering performance,robustness and stability.In addition,compared with other algorithms in the literature,ASFA achieves better effect in accuracy optimization aspect.关键词
萤火虫算法/K-means算法/初始聚类中心/自适应步长/鲁棒性Key words
firefly algorithm (FA)/K-means algorithm/initial clustering center/self-adaptive step length/robustness分类
信息技术与安全科学引用本文复制引用
潘晓英,陈雪静,李昂儒,赵普..基于自适应步长的萤火虫划分聚类算法[J].计算机应用研究,2017,34(12):3576-3579,3602,5.基金项目
国家自然科学基金资助项目(61105064,61203311) （61105064,61203311）
陕西省教育厅专项科研计划项目(14JK1665) （14JK1665）
厦门市科技计划项目(3502Z20141164) （3502Z20141164）
