摘要
有序样本的聚类关键是建立合理的分类指标。将系统聚类法与Fisher算法相结合,引入样本一阶自相关系数描述相邻样本间的相关关系,并以此为分类指标,进行分类。最后运用离差平方和确定最优分类数及分类结果,分类结果符合实际。
The key of cluster of ordered sample is establishing the reasonable classification index. Combining the system cluster method and Fisher algorithm ,establishing the first order self-correlation coefficient ,classifing the samples using it. Lastly, determining the best classification number and result.
出处
《廊坊师范学院学报（自然科学版）》
2013年第6期21-23,27,共4页
Journal of Langfang Normal University(Natural Science Edition)
关键词
有序样本
白相关系数
离差平方和
聚类
ordered sample
self-correlation coefficient
square deviation
cluster
