摘要
针对传统K均值聚类(K-Means Clustering,KMC)算法在对电影信息数据聚类的过程中,初始聚类中心选取随机性较大、聚类结果不稳定且算法容易陷入局部最优、影响迭代精度等不足,提出一种基于随机跳跃式翻滚觅食蝠鲼优化的K均值联合迭代聚类算法(MRRJRFO-KMC),实现对电影信息数据的聚类.首先,提出一种均值最大最小距离积法来初始化聚类中心,改善聚类中心选取的随机性,避免随机初始化对聚类结果造成的不稳定性.其次,在迭代的过程中加入蝠鲼觅食优化算法,并对蝠鲼觅食优化算法中螺旋觅食和翻滚觅食进行改进,提出一种随机跳跃式翻滚觅食蝠鲼优化的策略,解决了蝠鲼觅食优化算法易陷入局部最优的问题.将随机跳跃式翻滚觅食蝠鲼优化算法加入KMC算法,对KMC算法迭代过程中的聚类中心进行优化,提高了聚类精度.在Iris,Aggregation,Ecoli和Seeds国际标准数据集上对MRRJRFO-KMC算法、MRFO-KMC算法、KMC算法、K-Means++算法、模糊C均值(Fuzzy C-Means,FCM)聚类算法进行比较测试,实验结果表明,MRRJRFO-KMC算法和其他算法相比,准确性和收敛速度都有所提升.在电影信息数据处理过程中,该算法能够根据所给的信息进行有效的聚类,应用价值明显.
For the traditional K-Means Clustering(KMC)algorithm in the process of clustering movie information data,the selection of initial clustering center is relatively random,the clustering result is unstable and the algorithm is easy to fall into local optimum,affecting the iterative accuracy and other shortcomings.This paper proposes a K-Means joint iterative clustering algorithm based on Manta Ray with Random Jumping Roll Foraging Optimization algorithm(MRRJRFO-KMC)to realize the clustering of movie information data.Firstly,a mean max-min distance product method is proposed to initialize the cluster centers,which improves the randomness of cluster center selection and avoids the instability of clustering results caused by random initialization.Secondly,the Manta Ray Foraging Optimization algorithm is added in the iterative process,and spiral foraging and roll foraging in the Manta Ray Foraging Optimization algorithm are improved.A Manta Ray with Random Jumping Roll Foraging Optimization algorithm is proposed,which solves the problem that the Manta Ray Foraging Optimization algorithm is easy to fall into local optimum.The Manta Ray with Random Jumping Roll Foraging Optimization algorithm is added to KMC algorithm,and the clustering center in the KMC algorithm is optimized,which improves the clustering precision.MRRJRFO-KMC algorithm,MRFO-KMC algorithm,KMC algorithm,K-Means++algorithm and Fuzzy C-Means(FCM)algorithm are calculated on Iris,Aggregation,Ecoli and Seeds international standard datasets.Experimental results show that the accuracy and convergence speed of MRRJRFO-KMC algorithm are improved compared with other algorithms.In the process of movie information data processing,the algorithm can effectively cluster according to the given information,and its application value is obvious.
作者
黄鹤
李潇磊
王珺
王会峰
茹锋
Huang He;Li Xiaolei;Wang Jun;Wang Huifeng;Ru feng(Xi′an Key Laboratory of Intelligent Expressway Information Fusion and Control,Chang′an University,Xi′an,710064,China;School of Electronic and Control Engineering,Chang′an University,Xi′an,710064,China)
出处
《南京大学学报（自然科学版）》
CAS
CSCD
北大核心
2022年第5期856-867,共12页
Journal of Nanjing University（Natural Science）
基金
国家重点研发计划(2021YF2501200)
国家自然科学基金(52172379,52172324)
陕西省重点研发计划(2021SF-483)
陕西省自然科学基础研究计划(2021JM-184)
西安市智慧高速公路信息融合与控制重点实验室(长安大学)开放基金(300102321502)
中央高校基本科研业务费(300102240203)
作者简介
通讯联系人:王珺,E-mail:jwang@nwu.edu.cn
