节点文献
人工神经网络求一般对称矩阵特征向量的研究
Theoretical Feasibilty of Finding the Eigenvector of an Ormnary Symmetric Matrix Using an Artificlal Neural Network
【摘要】 讨论了对称矩阵特征向量的人工神经网络求法,特别是特征值重数大于1的情况。通过对已有人工神经网络数学模型的细致分析,得出了“只要输入电压适当,一般对称矩阵全部特征向量可由输出电压给出”的结论。
【Abstract】 A method is proposed for finding the eigenvector of any symmetric matrix using a neuralnetwork,especially for the case in which the degree of the eigenvalue is greater than 1.Through a careful analysis of the existing mathematical model for the artificial neural net-work,it has been found that all the eigenvectors of an ordinary symmetric matrix are givenby the stable voltage output if the voItage input is appropriate.
【基金】 国家自然科学基金
- 【文献出处】 华中理工大学学报 ,JOURNAL OF HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY , 编辑部邮箱 ,1995年08期
- 【分类号】TP183
- 【下载频次】59