基于遗传FCM算法的文本聚类

Text clustering based on genetic fuzzy C-means algorithm

提出了基于遗传FCM算法的文本聚类方法,首先采用LSI方法对文本特征进行降维,然后通过聚类有效性分析得到文本的类别数,最后采用遗传FCM算法对文本进行聚类。该方法较好地克服了FCM算法收敛于局部最优的缺陷,很好地解决了FCM算法对初始聚类中心敏感的问题。实验结果表明该方法具有较好的聚类性能。

A text clustering method based on genetic fuzzy c-means algorithm was proposed.At first,latent semantic index was used to reduce the dimension of text feature and then the number of text class was obtained through analyzing the validity of clustering.At last,genetic FCM algorithm was used to cluster the text.The proposed method overcomes the flaw of FCM algorithm which may converge to local optimum,and it resolves the problem of FCM algorithm which is sensitive to the initialized value of cluster center.The...