【作者】 贾文艳； 周萍； 邵颉； 欧阳楷；

【Author】 Jia Wenyan Zhou Ping Shao Jie Ou-yang Kai Department of BME,Capital University of Medical Sciececs,Beijing,100054

【机构】 首都医科大学生物医学工程系；

【摘要】 本文是符号序列处理的系列文章之一,叙述了用仿真方法对用于符号序列处理的神经网络模型的性质的研究,结果表明它与Kohonen的符号树方法相似,另外,由网络权重可以看出它与复杂度方法和符号统计方法是一致的。更多还原

【Abstract】 This paper is one of a series of papers about symbolic sequences,which describes a study of the property of the neural network used for symbolic sequence processing by simulation.The result showed that this method is similar to the symbol tree method proposed by Kohonen,and from the weights of the network, some property that coincides with those properties of complexity and symbol statistics can be obtained.更多还原