"基于WEB的普通话自动测评系统研究及实验结果分析"

基于WEB的普通话自动测评系统的研究，是南通大学毕业设计的主要内容。该系统的研究旨在有效地避免人工测试中的弊端，并节省大量的人力和资源。研究发现，通过采集大量的普通话语音信号并建立语音库，通过分析信号波形并进行比对匹配，可以实现语音识别和自动测评。然而，由于条件有限，本系统采用了微软提供的语音程序开发组件Speech SDK 5.1，通过调用相关的语音应用程序接口(SAPI)来实现普通话考试的自动测评。本论文的主要内容包括了普通话测试相关题库的建立、普通话语音的采集与分析、对采集的语音与标准答案进行比对算法的研究，以及计算机自动阅卷评分的研究。实验测试结果表明，该系统能够实现对单字、双字等项目的测试并评分，取得了预期的效果，为今后的进一步研究提供了良好的参考。 关键词：普通话，自动测评，语音识别，SPEECH SDKI ABSTRACT This paper studies the Mandarin automatic test system based on Web. The system will effectively avoid the shortcomings of manual testing and save a lot of manpower and resources. It is found that by collecting a large number of Mandarin speech signals and establishing a speech library, analyzing the signal waveform, and comparing and matching can achieve speech recognition and automatic evaluation. However, due to limited conditions, this system uses Microsoft's Speech SDK 5.1 to call the relevant speech application interface (SAPI) to achieve automatic evaluation of Mandarin tests. The main contents of this paper include the establishment of a related question bank for Mandarin tests, the collection and analysis of Mandarin speech, the research on the comparison algorithm between the collected speech and standard answers, and the research on computer automatic scoring. The experimental results show that the system can achieve testing and scoring for single and double characters, and has achieved the expected results, providing a good reference for future research. Keywords: Mandarin, automatic evaluation, speech recognition, SPEECH SDKI 根据所提供的内容，总结描述基于WEB的普通话自动测评系统的研究，重点包括研究背景、研究方法、实验结果以及对未来研究的展望，并对论文中所涉及到的关键词进行解释和说明。