id: 05811662
dt: j
an: 05811662
au: Zhu, Haodong; Zhong, Yong; Zhao, Xianghui
ti: An optimization initial center K-means algorithm for text clustering.
so: J. Zhengzhou Univ., Nat. Sci. Ed. 41, No. 2, 29-32 (2009).
py: 2009
pu: Zhengzhou University, Zhengzhou; International Book Trading Corporation,
    Beijing
la: ZH
cc: 
ut: K-means algorithm; simulated annealing algorithm; initial center
ci: 
li: 
ab: Summary: Owing to its random initial centers, unstable results are often
    obtained when using traditional K-means and its variants. A technique
    of optimizing initial centers of clustering is proposed. Combining an
    improved Simulated Annealing Algorithm (ISAA) with K-means, ISAA
    chooses initial clustering centers, and then an improved K-means
    (IK-means) is presented. So combining ISAA with K-means, IK-means has
    better advanced capability under either global or local conditions, and
    can resolve the problem that clustering results of K-means are likely
    influenced by initial centers. Through experiment testing, the accuracy
    and the stability of IK-means are proved to be better than those of
    K-means.
rv: