基于数据挖掘的普通话韵律规则学习
|
摘要
普通话韵律规则对于语音合成和语音学研究具有重要意义 .为了更有效地进行韵律规则学习 ,该文利用数据挖掘技术从语料库中提取规则 .通过聚类分析进行基频模式提取 ,并以此进行基频序列的离散化 ;由语言学分析的结果得出训练句子中每个音节的参数 ,利用决策树和神经网络学习音节的韵律变化规则 .测试表明基于数据挖掘的韵律规则学习取得了较好的结果 ,证实了方法的有效性 .
Mandarin prosodic models are very important in speech research and speech synthe sis, which mainly describess the variation of pitch. The models that are now being u sed in most Chinese Text\|To\|Speech systems are constructed by expert, qualitatively an d with low precision. In this paper, Data Mining is used to extract more accurate prosodic pattern s from actual large mandarin speech database to improve the naturalness and intelligibility of synth esized speech. In data preprocessing, typical prosody models are found by clustering analysis, a nd the original pitches extracted from sentences are discrete with classic pitch models. These clusters together with some linguistic features (including tone combination, word length, part\|of \|speech (POS), syllable position in word, word position in phrase) obtained by text parsing are use to acquire training data. ANN and Decision tree are trained respectively using above integr ated data to learn the variation prosody models of pitch. Two decisino trees are construc ted for predicting the classic pitch model and length of pitch based on C4.5, and BackPropagation(BP) network is used to learn the mapping between the linguistic features and the mean value of pit ch. Encouraging experimental results show the effectiveness of the proposed method base on Data Mining.
Mandarin prosodic models are very important in speech research and speech synthe sis, which mainly describess the variation of pitch. The models that are now being u sed in most Chinese Text\|To\|Speech systems are constructed by expert, qualitatively an d with low precision. In this paper, Data Mining is used to extract more accurate prosodic pattern s from actual large mandarin speech database to improve the naturalness and intelligibility of synth esized speech. In data preprocessing, typical prosody models are found by clustering analysis, a nd the original pitches extracted from sentences are discrete with classic pitch models. These clusters together with some linguistic features (including tone combination, word length, part\|of \|speech (POS), syllable position in word, word position in phrase) obtained by text parsing are use to acquire training data. ANN and Decision tree are trained respectively using above integr ated data to learn the variation prosody models of pitch. Two decisino trees are construc ted for predicting the classic pitch model and length of pitch based on C4.5, and BackPropagation(BP) network is used to learn the mapping between the linguistic features and the mean value of pit ch. Encouraging experimental results show the effectiveness of the proposed method base on Data Mining.
引文
1 L iu Qing-Feng,Ni Ji-Fu,Wang Ren-Hua. Research onimproving naturalness of synthesized specch.In:Proceedingsof 3rd China Computer Intelligence Interface and ApplicationWorkshop,1997. 16 3-16 8(in Chinese)(刘庆峰 ,倪晋富 ,王仁华 .提高合成语音自然度的研究 .见 :第3届中国计算机智能接口与智能应用学术会议论文集 ,1998.16 3-16 8)
2 Wu Zong-Ji. The tone variation in mandarin. ChineseGrammar,1982 ,(6 ) :439-4 4 9(in Chinese)(吴宗济 .普通话语句中的声调变化 .中国语文 . 1982 ,(6 ) :439-4 4 9)
3 L in Mao-Can,Yan Jin-Zu,Sun Guo-Hua. Experim ent of thenormal accent in Beijing dialect.Dialect,1984,(1) (in Chinese)(林茂灿 ,颜景助 ,孙国华 .北京话两字组正常重音的初步实验 .方言 ,1984,(1) )
4Fen L ong. The Duration of Vowel and Consonant forMandarin. Experim ental in Beijing Dialect. Beijing:PekingUniversity Press,1985 . 131-195 (in Chinese)(冯 隆 .北京话语流中声韵调的时长 .北京语音实验录 .北京 :北京大学出版社 ,1985 . 131-195 )
5 Usama M Fayyad,Gregory Piatetsky-Shapiro,Padhraic Smythet al. EDITORS,Adavance in Knowledge Dicovery and DataMining.AAAI/ MIT Press,1996
6 George H John. Enhancem ents to the data mining process[PhD dissertation] .Stanford U niversity,1997
7Davood Raflei,Alberto Mendelzon. Sim ilarity-based queriesfor tim e series data.In:Proceedings of the ACM SIGMODConference on the Management of Data (Sigmod'97) ,Tucson,Arizona,USA,1997. 13-2 5
8Wang Bi-Quan,Chen Zu-Yin.Pattern Recognization:Theory,Method and Application. Beijing:Earcth Quake Press,1989(inChinese)(王碧泉 ,陈祖荫 .模式识别 :理论、方法和应用 .北京 :地震出版社 ,1989)
9Quinlan J R. Induction of decision trees. Machine L earning,1986 ,1:81-10 6
10 Quinlan JR.C4.5 :Programs for Machine L earning.MorganKaufmanns Publishers,1993
11Hu Shou-Ren,Yu Shao-Bo,Dai Kui.Introduction to NeuralNetwork. Changsha:National University of DefendTechnology Press,1993(in Chinese)(胡守仁 ,余少波 ,戴 葵 .神经网络导论 .长沙 :国防科技大学出版社 ,1993)
12 Wang Wei. Principle of Artificial Neural Netowrk——Rudiment and Implem ent. Beijing:University of Aeronauticsand Astronautics Press,1995 (in Chinese)(王 伟 .人工神经网络原理——入门与应用 .北京 :北京航空航天大学出版社 ,1995 )
2 Wu Zong-Ji. The tone variation in mandarin. ChineseGrammar,1982 ,(6 ) :439-4 4 9(in Chinese)(吴宗济 .普通话语句中的声调变化 .中国语文 . 1982 ,(6 ) :439-4 4 9)
3 L in Mao-Can,Yan Jin-Zu,Sun Guo-Hua. Experim ent of thenormal accent in Beijing dialect.Dialect,1984,(1) (in Chinese)(林茂灿 ,颜景助 ,孙国华 .北京话两字组正常重音的初步实验 .方言 ,1984,(1) )
4Fen L ong. The Duration of Vowel and Consonant forMandarin. Experim ental in Beijing Dialect. Beijing:PekingUniversity Press,1985 . 131-195 (in Chinese)(冯 隆 .北京话语流中声韵调的时长 .北京语音实验录 .北京 :北京大学出版社 ,1985 . 131-195 )
5 Usama M Fayyad,Gregory Piatetsky-Shapiro,Padhraic Smythet al. EDITORS,Adavance in Knowledge Dicovery and DataMining.AAAI/ MIT Press,1996
6 George H John. Enhancem ents to the data mining process[PhD dissertation] .Stanford U niversity,1997
7Davood Raflei,Alberto Mendelzon. Sim ilarity-based queriesfor tim e series data.In:Proceedings of the ACM SIGMODConference on the Management of Data (Sigmod'97) ,Tucson,Arizona,USA,1997. 13-2 5
8Wang Bi-Quan,Chen Zu-Yin.Pattern Recognization:Theory,Method and Application. Beijing:Earcth Quake Press,1989(inChinese)(王碧泉 ,陈祖荫 .模式识别 :理论、方法和应用 .北京 :地震出版社 ,1989)
9Quinlan J R. Induction of decision trees. Machine L earning,1986 ,1:81-10 6
10 Quinlan JR.C4.5 :Programs for Machine L earning.MorganKaufmanns Publishers,1993
11Hu Shou-Ren,Yu Shao-Bo,Dai Kui.Introduction to NeuralNetwork. Changsha:National University of DefendTechnology Press,1993(in Chinese)(胡守仁 ,余少波 ,戴 葵 .神经网络导论 .长沙 :国防科技大学出版社 ,1993)
12 Wang Wei. Principle of Artificial Neural Netowrk——Rudiment and Implem ent. Beijing:University of Aeronauticsand Astronautics Press,1995 (in Chinese)(王 伟 .人工神经网络原理——入门与应用 .北京 :北京航空航天大学出版社 ,1995 )
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心