摘要
为了改进协同聚类中计算量较大的问题,提出一种信息熵加权的模糊协同聚类算法。首先引入信息熵来衡量隶属度差异矩阵中包含的不确定性信息,然后根据有效信息量定义相似性距离中的权重,最后通过权重对聚类的贡献实现子集之间的协同聚类。实验结果显示,新算法能充分利用数据子集中蕴涵的相关信息,以较高的计算效率实现更准确的协同聚类。与已有算法相比,新算法能自适应地计算协同关系强度,简化了参数设置和协同函数的复杂计算。
In order to overcome the disadvantage of large amount of calculation in collaborative clustering,this paper proposed an entropy-weighted fuzzy collaborative clustering algorithm. First,it introduced the entropy to measure the uncertain information in the difference matrices of membership degree. Second,it defined the entropy-weighted distance for similarity according to the available information. Last,it offered and implemented the collaborative clustering by means of the contribution of the weighted index. The experimental results show that the improved algorithm can take full advantage of the relevant information among the subsets and achieve more accurate collaborative results with high computational efficiency. Compared with the existing clustering,the improved algorithm can automatically calculate the collaborative relationship strength,so it can simplify the assignment of parameters and the calculation of collaborative function.
出处
《计算机应用研究》
CSCD
北大核心
2015年第4期1016-1018,1023,共4页
Application Research of Computers
基金
国家自然科学基金青年基金资助项目(61402202)
高等学校博士学科点专项科研基金资助项目(20120093120016)
江苏省自然科学青年基金资助项目(BK20130117)
关键词
模糊聚类
协同关系
差异矩阵
信息熵
权重系数
fuzzy clustering
collaborative relationship
difference matrix
information entropy
weight coefficient
