中文摘要：

提出了一种新的动态模糊自组织神经网络模型（TGFCM）,并将其用于文本聚类中。针对传统模糊自组织神经网络需要预先确定聚类数的问题,TGFCM采用了可自动确定聚类数的动态自组织神经网络（TGSOM）的结构,在TGSOM网络结构中提出新的学习率计算式,并以模糊聚类中心作为TGFCM网络中对应的神经元的权值,从而提高了聚类的精度,并可提高收敛速度。

英文摘要：

This paper proposed a new model of dynamic fuzzy Kohonen neural network（TGFCM）,which was applied to the text clustering.TGFCM adopted the structure of the dynamics self-organization maps（TGSOM） which could determine the number of clustering automatically.Proposed a new calculation formula of the learning rate of TGSOM in the TGFCM.TGFCM used fuzzy clustering central vectors as the corresponding neuron weights.Both improved the precision of clustering and the rate of convergent of the network.