论文部分内容阅读
聚类分析神经网络的输入样本序列具有各种不同的空间分布性态，这就要求采用不同的相似匹配准则。本文以广义距离和广义相似匹配准则为基本概念，介绍了广义聚类神经网络ＧＣ的设计思想和网络的学习训练方法，该方法具有广泛的适用性。
Clustering neural network input sample sequence has a variety of different spatial distribution, which requires the use of different similar matching criteria. In this paper, the generalized distance and generalized similarity matching criteria are taken as the basic concepts. The design idea of GCNN and the learning and training method of network are introduced. The method has wide applicability.