关键词: 分类； 模糊C均值； kNN划分； 新的模糊C均值算法

Abstract:

Two methods for initialization of cluster centers were presented, one is supervised method; the other is the method based on kNN partition. An improved algorithm based on fuzzy Cmeans, a traditional clustering algorithm was finished, which was used to complete the data classification first. And several standard datasets from UCI were tested, which shows a well classification result is found.

Key words: classification, fuzzy Cmeans, kNN partition, NFCM