基于神经网络的聚类方法研究

Research on Clustering Method Based on Neural Network

针对常用聚类方法不能有效处理噪声数据的问题,本文结合神经网络具有自适应性的特点,提出基于神经网络的聚类(NN_Cluster)模型,并设计了基于自适应共振理论的神经网络聚类模型(ARTNN_Cluster)和基于自组织特征映射的神经网络聚类模型(SOMNN_Cluster)。标准数据集上的实验结果表明,与传统的K_means聚类方法相比,本文提出的基于神经网络的聚类模型有效地克服了传统方法的噪声问题,得到了较好的聚类效果。

This paper presents a clustering model based on neural network（NN_Cluster） combining the self adaptive feature of neural network,in order to solve the noise data of clustering.Then design two clustering algorithms based on adaptive resonance theory neural network（ARTNN_Cluster） and self-organizing feature map neural networks（SOMNN_Cluster）.Simulation results on UCI datasets demonstrate that comparing with traditional K_means clustering means,the NN_Cluster effectively overcome the noise of traditional clustering methods and the better clustering results are obtained by this model.