摘要
聚类分析是数据挖掘的重要技术,是一种无监督的学习方式,可根据数据间的相似程度,将数据进行分类.竞争决策算法是一种基于竞争造就优化和决策左右结果的新型优化算法,针对聚类分析的特点,设计了一种竞争决策算法进行求解,经实验测试和验证,并与其它算法的结果进行比较,获得了较好的结果.
Clustering analysis is an important technique in data mining. It is an unsupervised learning technique and is a division of data into groups of similar objects. A competitive decision algorithm is a new optimization algorithm based on the characteristics that competition builds optimization and the result of competition hinges on decision. Based on some characteristics of clustering analysis, this paper provides a competitive decision algorithm for clustering analysis. The experiments are tested and the results show that the clustering result of competitive decision algorithm is better than that of some other algorithms.
出处
《数学的实践与认识》
CSCD
北大核心
2008年第21期58-62,共5页
Mathematics in Practice and Theory
基金
国家自然科学基金项目资助(70871081)
上海市重点学科建设项目资助(T0502)
上海市高校选拔培养优秀青年教师科研专项基金项目资助(21012)
关键词
聚类分析
竞争力函数
决策函数
竞争决策均衡
竞争决策算法
clustering analysis
competitive force function
decision function
competitive decision equilibrium
competitive decision algorithm