摘要
在传统的K-means算法中,K值和初始聚类中心往往凭人的经验或随机选取,算法对选取结果又比较敏感,同时算法易陷入局部最优。论文针对这些不足,利用遗传算法的全局寻优特性和自适应搜索概率技术等优势,改善K-means聚类方式。仿真实验表明,新算法在平均迭代次数和准确率方面优于传统K-means算法。
In the traditional K-means algorithm,the K value and the initial clustering center are often selected by human ex⁃perience or random.The algorithm is sensitive to the selection result,and the algorithm is easy to fall into local optimum.In view of these shortcomings,this paper uses the advantages of genetic algorithm global optimization and adaptive search probability technolo⁃gy to improve the K-means clustering method.Simulation experiments show that the new algorithm is superior to the traditional K-means algorithm in terms of average iteration number and accuracy.
作者
冯永亮
李浩
FENG Yongliang;LI Hao(School of Information Engineering,Xi'an University,Xi'an 710065;Xi'an Internet of Things Application Engineering Laboratory,Xi'an University,Xi'an 710065)
出处
《计算机与数字工程》
2020年第8期1831-1834,1839,共5页
Computer & Digital Engineering
基金
陕西省自然科学基金项目(编号:2018JM6100)
陕西省教育厅科学研究计划项目(编号:18JK1149)
西安市科技计划重点项目(编号:2017CGWL13)资助。
