基于BiRNN的维吾尔语情感韵律短语注意力模型
|
摘要
当前维吾尔语情感语音合成采用韵律边界预测方法来实现情感语音转换。通过该方法合成出来的语音,虽然可表现出相应的情感,然而其情感表现力不够理想。针对此问题,该文提出一种基于BiRNN的维吾尔语情感韵律短语注意力模型。在情感韵律转换前使用该模型进行情感分类,并将其分类结果作为韵律边界预测的输入,改进了情感韵律转换方法。使用改进的词性特征向量和韵律短语向量作为词向量的补充,从而有效提升维吾尔文文本情感分类的准确率。实验结果表明,该模型由两个单词构成的韵律短语作为特征时,准确率在维吾尔五分类情感数据集上达到了很好的分类效果。
At present, Uyghur sentimental speech synthesis uses prosodic boundary prediction method to realize emotional speech conversion. The speech synthesized by this method can express the corresponding emotions, but its emotional expression is not ideal. To solve this problem, this paper proposes an attention model of Uygur emotional prosodic phrases based on BiRNN. The model is used to classify emotion before prosodic conversion, and the classification results are used as input for prosodic boundary prediction to improve the method of prosodic conversion. The improved part-of-speech feature vector and prosodic phrase vectors are used to supplement the word vector, which effectively improve the accuracy of Uyghur text sentiment classification. The experimental results show that when the prosodic phrase composed of two words is used as a feature, the accuracy of the model achieves the best classification effect on the Uyghur five-category sentiment data set.
At present, Uyghur sentimental speech synthesis uses prosodic boundary prediction method to realize emotional speech conversion. The speech synthesized by this method can express the corresponding emotions, but its emotional expression is not ideal. To solve this problem, this paper proposes an attention model of Uygur emotional prosodic phrases based on BiRNN. The model is used to classify emotion before prosodic conversion, and the classification results are used as input for prosodic boundary prediction to improve the method of prosodic conversion. The improved part-of-speech feature vector and prosodic phrase vectors are used to supplement the word vector, which effectively improve the accuracy of Uyghur text sentiment classification. The experimental results show that when the prosodic phrase composed of two words is used as a feature, the accuracy of the model achieves the best classification effect on the Uyghur five-category sentiment data set.
引文
[1]帕丽旦·木合塔尔,吾守尔·斯拉木.基于HMM的维吾尔语可训练语音合成文本分析关键技术研究[D].乌鲁木齐:新疆大学,2012.PALIDAN Muhetaer,WUSHOUER Silamu.Research the text analysis key technologies of Uyghur trainable synthetic system based on HMM[D].Urumqi:Xinjiang University,2012.
[2]HU M,LIU B.Mining and summarizing customer reviews[C]//Proceedings of the 10th ACM SIGKDDInternational Conference on Knowledge Discovery and Data Mining.Seattle:ACM,2004:168-177.
[3]KIM S M,HOVY E.Determining the sentiment of opinions[C]//International Conference on Computational Linguistics.[S.l.]:Association for Computational Linguistics,2004:1367.
[4]KAMPS J.Using wordnet to measure semantic orientation of adjectives[C]//International Conference on Language Resources and Evaluation.Centro Cultural De Belem,Lisbon,Portugal:[s.n.],2004:1115-1118.
[5]朱嫣岚,闵锦,周雅倩,等.基于How Net的词汇语义倾向计算[J].中文信息学报,2006,20(1):14-20.ZHU Yan-lan,MIN Jin,ZHOU Ya-qian,et al.Semantic orientation computing based on hownet[J].Journal of Chinese Information Processing,2006,20(1):14-20.
[6]KENNEDY A,INKPEN D.Sentiment classification of movie reviews using contextual valence shifters[J].Computational Intelligence,2006,22(2):110-125.
[7]KEVIN C,VIDYASAGAR P,THARAMS D.General inquirer categories[EB/OL].[2017-05-15].http://www.wjh.harvard.edu/~inquirer/homecat.htm.
[8]OHANA B,TIERNEY B.Sentiment classification of reviews using sentiment word net[C]//IT&T Conference.[S.l.]:Springer,2009:10.21427/D77S56.
[9]TURNEY P D.Thumbs up or thumbs down?Semantic orientation applied to unsupervised classification of reviews[C]//Meeting on Association for Computational Linguistics.[S.l.]:Association for Computational Linguistics,2002:417.
[10]READ J,CARROLL J.Weakly supervised techniques for domain-independent sentiment classification[C]//International CIKM Workshop on Topic-Sentiment Analysis for MASS Opinion.[S.l.]:ACM,2009:45-52.
[11]PANG B,LEE L,VAITHYANATHAN S.Thumbs up?Sentiment classification using machine learning techniques[C]//In ACL-02 Conference on Empirical Methods in Natural Language Processing.[S.l.]:Association for Computational Linguistics,2002:79-86.
[12]SOCHER R,PENNINGTON J,HUANG E H,et al.Semi-supervised recursive autoencoders for predicting sentiment distributions[C]//Conference on Empirical Methods in Natural Language Processing.Edinburgh,UK:[s.n.],2011:151-161.
[13]朱少杰.基于深度学习的文本情感分类研究[D].哈尔滨:哈尔滨工业大学,2014.ZHU Shao-jie.Research on text sentiment classification based on deep learning[D].Harbin:Harbin University of Technology,2014.
[14]KIM Y.Convolutional neural networks for sentence classification[EB/OL].[2017-09-20].https://arxiv.org/abs/1408.5882.
[15]NARENDRA B,SAI K U,RAJESH G,et al.Sentiment analysis on movie reviews:a comparative study of machine learning algorithms and open source technologies[J].International Journal of Intelligent Systems Technologies and Applications,2016,8(8):66-70.
[16]SEVERYN A,MOSCHITTI A.Twitter sentiment analysis with deep convolutional neural networks[C]//International ACM SIGIR Conference.[S.l.]:ACM,2015:10.1145/2766462.2767830.
[17]SANTOS C N D,GATTIT M.Deep convolutional neural networks for sentiment analysis of short texts[C]//International Conference on Computational Linguistics.Dublin,Ireland:[s.n.],2014:69-78.
[18]SOCHER R,HUVAL B,MANNING D,et al.Semantic compositionality through recursive matrix-vector spaces[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing.Jeju Island,Korea:[s.n.],2012:1201-1211.
[19]王树恒,吐尔根·依布拉音,卡哈尔江·阿比的热西提,等.基于BLSTM的维吾尔语文本情感分析[J].计算机工程与设计,2017(10):2879-2886.WANG Shu-heng,TURGUN Ibrahim,Kahaerjiang Abiderexiti,et al.Sentiment classification of Uyghur text based on BLSTM[J].Computer Engineering and Design,2017(10):2879-2886.
[20]王荣波,谌志群,周建政,等.基于Wikipedia的短文本语义相关度计算方法[J].计算机应用与软件,2015,32:82-86.WANG Rong-bo,CHEN Zhi-qun,ZHOU Jian-zheng,et al.Short texts semantic relevance computation method based on wikipedia[J].Computer Applications and Software,2015,32:82-86.
[21]HOCHREITER S,SCHMIDHUBER J.Long short-term memory[J].Neural Computation,1997,9:1735-1780.
[22]ALEX G,GREG W,DANIHELKA I,et al.Neural turing machines[EB/OL].[2017-10-17].https://arxiv.org/abs/1410.5401v2.
[23]OUARDA W,TRICHILI H,ALIMI A M,et al.MLPneural network for face recognition based on gabor features and dimensionality reduction techniques[C]//Multimedia computing and systems(ICMCS).Marrakech,Morocco:[s.n.],2014:10.1109/ICMCS.2014.6911265.
[24]LI Dong,WEI Fu-ru,MING Zhou,et al.Question answering over freebase with multicolumn convolutional neural networks[C]//Proc of the 53rd Annual Meeting of the Association for computational Linguistics and the 7th International Joint Conference on Natural Language Processing.Beijing:[s.n.],2015:10.3115/v1/p15-1026.
[25]何炎祥,孙松涛,牛菲菲,等.用于微博情感分析的一种情感语义增强的深度学习模型[J].计算机学报,2017,40(4):773-790.HE Yan-xiang,SUN Song-tao,NIU Fei-fei,et al.A deep learning model enhanced with emotion semantic microblog sentiment analysis[J].Chinese Journal of Computers,2017,40(4):773-790.
[27]李松如,陈锻生.采用循环神经网络的情感分析注意力模型[J].华侨大学学报(自然科学版),2018(2):252-255.LI Song-ru,CHENG Duan-sheng,Recurrent neural network using attention model for sentiment analysis[J].Journal of Huaqiao University(Natural Science),2018(2):252-255.
[28]艾力·海如拉,吾守尔·斯拉木.基于神经网络的维吾尔文词向量表示方法及其应用研究[D].乌鲁木齐:新疆大学,2018.AILI Hairula,WUSHOUER Silamu.Research on neural networks based Uyghur word vectors representation and its application[D].Urumqi:Xinjiang University,2018.
[2]HU M,LIU B.Mining and summarizing customer reviews[C]//Proceedings of the 10th ACM SIGKDDInternational Conference on Knowledge Discovery and Data Mining.Seattle:ACM,2004:168-177.
[3]KIM S M,HOVY E.Determining the sentiment of opinions[C]//International Conference on Computational Linguistics.[S.l.]:Association for Computational Linguistics,2004:1367.
[4]KAMPS J.Using wordnet to measure semantic orientation of adjectives[C]//International Conference on Language Resources and Evaluation.Centro Cultural De Belem,Lisbon,Portugal:[s.n.],2004:1115-1118.
[5]朱嫣岚,闵锦,周雅倩,等.基于How Net的词汇语义倾向计算[J].中文信息学报,2006,20(1):14-20.ZHU Yan-lan,MIN Jin,ZHOU Ya-qian,et al.Semantic orientation computing based on hownet[J].Journal of Chinese Information Processing,2006,20(1):14-20.
[6]KENNEDY A,INKPEN D.Sentiment classification of movie reviews using contextual valence shifters[J].Computational Intelligence,2006,22(2):110-125.
[7]KEVIN C,VIDYASAGAR P,THARAMS D.General inquirer categories[EB/OL].[2017-05-15].http://www.wjh.harvard.edu/~inquirer/homecat.htm.
[8]OHANA B,TIERNEY B.Sentiment classification of reviews using sentiment word net[C]//IT&T Conference.[S.l.]:Springer,2009:10.21427/D77S56.
[9]TURNEY P D.Thumbs up or thumbs down?Semantic orientation applied to unsupervised classification of reviews[C]//Meeting on Association for Computational Linguistics.[S.l.]:Association for Computational Linguistics,2002:417.
[10]READ J,CARROLL J.Weakly supervised techniques for domain-independent sentiment classification[C]//International CIKM Workshop on Topic-Sentiment Analysis for MASS Opinion.[S.l.]:ACM,2009:45-52.
[11]PANG B,LEE L,VAITHYANATHAN S.Thumbs up?Sentiment classification using machine learning techniques[C]//In ACL-02 Conference on Empirical Methods in Natural Language Processing.[S.l.]:Association for Computational Linguistics,2002:79-86.
[12]SOCHER R,PENNINGTON J,HUANG E H,et al.Semi-supervised recursive autoencoders for predicting sentiment distributions[C]//Conference on Empirical Methods in Natural Language Processing.Edinburgh,UK:[s.n.],2011:151-161.
[13]朱少杰.基于深度学习的文本情感分类研究[D].哈尔滨:哈尔滨工业大学,2014.ZHU Shao-jie.Research on text sentiment classification based on deep learning[D].Harbin:Harbin University of Technology,2014.
[14]KIM Y.Convolutional neural networks for sentence classification[EB/OL].[2017-09-20].https://arxiv.org/abs/1408.5882.
[15]NARENDRA B,SAI K U,RAJESH G,et al.Sentiment analysis on movie reviews:a comparative study of machine learning algorithms and open source technologies[J].International Journal of Intelligent Systems Technologies and Applications,2016,8(8):66-70.
[16]SEVERYN A,MOSCHITTI A.Twitter sentiment analysis with deep convolutional neural networks[C]//International ACM SIGIR Conference.[S.l.]:ACM,2015:10.1145/2766462.2767830.
[17]SANTOS C N D,GATTIT M.Deep convolutional neural networks for sentiment analysis of short texts[C]//International Conference on Computational Linguistics.Dublin,Ireland:[s.n.],2014:69-78.
[18]SOCHER R,HUVAL B,MANNING D,et al.Semantic compositionality through recursive matrix-vector spaces[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing.Jeju Island,Korea:[s.n.],2012:1201-1211.
[19]王树恒,吐尔根·依布拉音,卡哈尔江·阿比的热西提,等.基于BLSTM的维吾尔语文本情感分析[J].计算机工程与设计,2017(10):2879-2886.WANG Shu-heng,TURGUN Ibrahim,Kahaerjiang Abiderexiti,et al.Sentiment classification of Uyghur text based on BLSTM[J].Computer Engineering and Design,2017(10):2879-2886.
[20]王荣波,谌志群,周建政,等.基于Wikipedia的短文本语义相关度计算方法[J].计算机应用与软件,2015,32:82-86.WANG Rong-bo,CHEN Zhi-qun,ZHOU Jian-zheng,et al.Short texts semantic relevance computation method based on wikipedia[J].Computer Applications and Software,2015,32:82-86.
[21]HOCHREITER S,SCHMIDHUBER J.Long short-term memory[J].Neural Computation,1997,9:1735-1780.
[22]ALEX G,GREG W,DANIHELKA I,et al.Neural turing machines[EB/OL].[2017-10-17].https://arxiv.org/abs/1410.5401v2.
[23]OUARDA W,TRICHILI H,ALIMI A M,et al.MLPneural network for face recognition based on gabor features and dimensionality reduction techniques[C]//Multimedia computing and systems(ICMCS).Marrakech,Morocco:[s.n.],2014:10.1109/ICMCS.2014.6911265.
[24]LI Dong,WEI Fu-ru,MING Zhou,et al.Question answering over freebase with multicolumn convolutional neural networks[C]//Proc of the 53rd Annual Meeting of the Association for computational Linguistics and the 7th International Joint Conference on Natural Language Processing.Beijing:[s.n.],2015:10.3115/v1/p15-1026.
[25]何炎祥,孙松涛,牛菲菲,等.用于微博情感分析的一种情感语义增强的深度学习模型[J].计算机学报,2017,40(4):773-790.HE Yan-xiang,SUN Song-tao,NIU Fei-fei,et al.A deep learning model enhanced with emotion semantic microblog sentiment analysis[J].Chinese Journal of Computers,2017,40(4):773-790.
[27]李松如,陈锻生.采用循环神经网络的情感分析注意力模型[J].华侨大学学报(自然科学版),2018(2):252-255.LI Song-ru,CHENG Duan-sheng,Recurrent neural network using attention model for sentiment analysis[J].Journal of Huaqiao University(Natural Science),2018(2):252-255.
[28]艾力·海如拉,吾守尔·斯拉木.基于神经网络的维吾尔文词向量表示方法及其应用研究[D].乌鲁木齐:新疆大学,2018.AILI Hairula,WUSHOUER Silamu.Research on neural networks based Uyghur word vectors representation and its application[D].Urumqi:Xinjiang University,2018.