基于粗糙集理论和SOFM神经网络的聚类方法

A CLUSTERING METHOD BASED ON ROUGH SETS THEORY AND SOFM NEURAL NETWORK

粗糙集理论和自组织特征映射SOFM(Self-Organizing-Feature-Map)神经网络在聚类分析中有各自的优势和劣势,结合SOFM神经网络和粗糙集理论提出一种算法。该算法利用粗糙集理论的属性约简去掉样本的冗余属性,并将处理过的数据作为SOFM神经网络的训练样本,从而减小了SOFM网络的规模,因此提高了样本的聚类效率。

According to the advantages and the disadvantages the Rough sets theory and the SOFM （self-organizing feature map） neural network have respectively, an algorithm is presented based on the combination of them. This algorithm reduces the redundant attribute of samples by using Rough sets theory, and then transfers the processed data to the SOFM neural network as its training data, so the scale of SOFM neural network is simplified, and the clustering efficiency has got improved.