节点文献
基于LVQ神经网络的混沌时间序列分类识别
Classification of Chaotic Time Series Using Neural Network
【摘要】 学习向量量化 (L VQ)是一种自适应数据分类方法 ,文中研究了利用这种神经网络对 Jeffcott转子碰摩模型的非线性混沌时间序列进行分类识别 ,得到了满意的效果。分析结果表明 ,该方法可以实现对这类混沌信号和其它响应信号数据的聚类 ,对非线性信号分类识别提供了一种较为直接的处理方法
【Abstract】 The idea of learning vector quantization (LVQ) is to find a natural grouping in a set of data. Every data vector is associated wi th a point in a d-dimensional data space and the vectors of the same class form a cluster in data space. In this paper, we try to classify the chaotic signals of the rotor-to-stator rub-impacting system using LVQ neural network. Such no nlinear signal can be classified directly without detecting fault characteristic s in advance. The results demonstrate that the LVQ neural network can cluster no nlinear chaotic time series and supply a direct method for classifying such nonl inear signals.
- 【文献出处】 机械科学与技术 ,Mechanical Science and Technology , 编辑部邮箱 ,2001年06期
- 【分类号】TP183
- 【被引频次】13
- 【下载频次】203