论文部分内容阅读
该文提出了一个动态的生长型树型自组织特征映射网络,用以模拟英语母语者汉语声调的认知发展。该网络既克服了传统的Kohonen自组织特征映射网络的固定网络结构限制以及容量有限性,又克服了其他类似动态网络的较弱的拓扑映射特性,可以较好地模拟英语母语者汉语声调认知的发展。模拟结果跟实验结果呈现出非常好的一致性,既证明了行为实验中汉语声调的动态发展过程,也为汉语声调认知的机制研究提供了机理上的解释。
This paper presents a dynamic growing tree of self-organizing feature maps to simulate the cognitive development of Chinese tone in English native speakers. The network not only overcomes the fixed network structure limitations and the limited capacity of the traditional Kohonen self-organizing feature mapping network, but also overcomes the weak topological mapping characteristics of other similar dynamic networks and can better simulate the Chinese tone recognition of English native speaker development of. The simulation results show very good agreement with the experimental results, which not only proves the dynamic development process of Chinese tone in behavioral experiments, but also provides a mechanistic explanation for the mechanism of Chinese tone recognition.