摘要
针对核模糊C-均值聚类算法中隶属度的计算特点,提出了一种改进的核模糊C-均值算法。改进后的算法是,在更新对象类的隶属度之前先判断对象是否可能属于该类。如果对象可能属于该类,则为其分配一个大于0的隶属度,否则直接将其隶属度置为0。针对不同测试数据集的实验结果表明,改进后的核模糊C-均值算法提高了聚类效果,是一种可行有效的算法。
According to the calculation feature of membership degree in fuzzy kernel c-means algorithm, an improved fuzzy kernel c-means algorithm is proposed. The new algorithm judges whether it is possible for the object to belong to the class before updating the membership degree. If the object is likely to belong to the class, the membership degree will be set as a positive number. Otherwise, the membership degree should be zero. The experiments show that the new algorithm improves the clustering result and it is a feasible and effective algorithm.
出处
《软件导刊》
2016年第1期40-42,共3页
Software Guide
关键词
模糊C-均值
聚类算法
隶属度
Fuzzy Kernel C-means
Clustering Algorithm
Membership Degree
