关键词: 声调语言, 声调发声, 声调识别, 汉语声调, 模式识别

Abstract: Objective Traditionally, tone production is evaluated subjectively using human listeners. The present study was designed to investigate the efficacy of using an artificial neural network in evaluating tone production of Mandarin-speaking children. Methods The subject group included 61 normal-hearing children aged between 3 and 9 years old. The fundamental frequency(F0) of their produced monosyllabic words was extracted. The F0 were then used as inputs to a feed-forward backpropagation artificial neural network. The numbers of inputs and neurons in the hidden layer were 12 and 16, respectively. The output layer consisted of 4 neurons representing the 4 Mandarin tone patterns. The tone-recognition performance of the neural network was further compared with that of native-Mandarin-speaking adult listeners. Results The neural network successfully classified the tone patterns of the 61 children speakers with an accuracy of about 85% correct. The score was shown to be significantly better than the perception score by the adult listeners. There was individual variability in the children's tone production accuracy as revealed by both the tone recognition of the neural network and by the tone perception of the adult listeners. Conclusion This study demonstrates that the artificial neural network can successfully classify Mandarin-Chinese tone patterns produced by multiple children. The neural network can be used as an objective way of evaluating tone production of children.

Key words: tone language, tone production, tone recognition, Mandarin tones, pattern recognition