聚类是数据挖掘中很重要的一部分.提出一种新的算法,不仅能避免最终的结果陷入局部解而且不用预知类别就可以对大批的数据进行分类,同时可以很容易的找出噪声数据.实验证明了这种算法在科学数据挖掘中是很有效的.
A new clustering algorithm is presented. It not only avoids the local solution in most cases, but also can classify a large quantity of data without knowing the classes beforehand and can distinguish the abnormal data. The experiment results show that the algorithm is very efficient in the scientific data mining.