請用此 Handle URI 來引用此文件：
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/63278完整後設資料紀錄
| DC 欄位 | 值 | 語言 |
|---|---|---|
| dc.contributor.advisor | 李琳山(Lin-Shan Lee) | |
| dc.contributor.author | Yu-Cheng Liu | en |
| dc.contributor.author | 劉又誠 | zh_TW |
| dc.date.accessioned | 2021-06-16T16:32:18Z | - |
| dc.date.available | 2013-01-16 | |
| dc.date.copyright | 2013-01-16 | |
| dc.date.issued | 2012 | |
| dc.date.submitted | 2012-12-05 | |
| dc.identifier.citation | [1] MIT OpenCourseWare, http://ocw.mit.edu/index.htm
[2] Hung-Yi Lee, Lin-Shan Lee, 'Improved Lattice-Based Spoken Document Retrieval by Directly Learning form the Evaluation Measures', ICASSP 2009 [3] J. Zhang, H.-Y. Chan, P. Fung, L. Cao, 'A Comparative Study on speech Summarization of Broadcast News and Lecture Speech', Interspeech 2007 [4] Yun-Nung Chen, Yu Huang, Sheng-Yi Kong, Lin-Shan Lee, 'Automatic Key Term Extraction from Spoken Course Lectures using Branching Entropy and Prosodic/Semantic Features', SLT 2010 [5] Daniel Jurafsky, James H. Martin, 'Speech and Language Processing', Second Edition [6] M. Zimmerman, D. Hakkani-Tur, J. Fung, N. Firghafori, L. Gottlieb, E. Shriberg, Y. Liu, “The ICSI+ Multilingual Sentence Segmentation System”, Interspeech 2006 [7] Chien-Yu Chou, Lin-Shan Lee, 'Chinese Sentence Segmentation using Machine Learning Methods', NTU Master Thesis 2009 [8] Shou-Chieh Hsu, Lin-Shan Lee, 'Topic Segmentation on Lecture Corpus and Its Application', NTU Master Thesis 2008 [9] S. Cuendet, E. Shriberg, B. Favre, J. Fung, D. Hakkani-Tur, 'An Analysis of Sentence Segmenation Features for Broadcast News, Broadcast Conversations, and Meetings', SIGIR 2007 [10] Che-Kuang Lin, Lin-Shan Lee, “New Approaches for Detecting Edit Disfluencies in Transcribing Spontaneous Mandarin Speech”, NTU Ph.D. Thesis 2009 [11] A. Stolcke, E. Shriberg, 'Automatic Linguistic Segmentation of Conversational Speech', ICSLP 1996 [12] E. Shriberg, A. Stolcke, D. Hakkani-Tur, Gokhan Tur, “Prosody-based Automatic Segmentation of Speech into Sentences and Topics”, Speech Communication 2000 [13] M. Johnson, 'PCFG Models of Linguistic Tree Representations', Computational Linguistics, vol. 24, no. 4, pp. 613-632, 1998 [14] Benoit Favre, Dilek Hakkani-Tur, Slav Petrov, Dan Klein, “Efficient Sentence Segmentation Using Syntactic Feature”, SLT 2008 [15] Yaser S. Abu-Mostafa, Malik Magdon-Ismail, Hsuan-Tien Lin, 'Learning From Data', 2012 [16] M. Magimai.-Doss, D. Hakkani-Tur, O. Cetin, E. Shriberg, J. Fung, N. Mirghafori, “Entropy Based Classifier Combination for Sentence Segmentation”, ICASSP 2007 [17] Li-Wei Cheng, Lin-Shan Lee, 'Prosody and Tone Modeling for Mandarin Chinese with Applications in Speech Recognition and Prosody Prediction', NTU Master Thesis 2008 [18] R.O. Duda, P.E. Hart, D.G. Stock, 'Pattern Classification', Second Edition, 2001 [19] Umit Guz, Sebastien Cuendet, Dilek Hakkani-Tur, Gokhan Tur, “Multi-View Semi-Supervised Learning for Dialog Act Segmentation of Speech”, IEEE Transactions on Audio, Speech, and Language Processing 2010 [20] Umit Guz, Sebastien Cuendet, Dilek Hakkani-Tur, Gokhan Tur, “Co-training Using Prosodic and Lexical Information for Sentence Segmentation”, Interspeech 2006 [21] Chih-Chung Chang and Chih-Jen Lin, 'LIBSVM：a library for support vector machines', ACM Transactions on Intelligent Systems and Technology, 2:27:1--27:27,2011. Software available at http://www.csie.ntu.edu.tw/~cjlin/libsvm [22] Andreas Stolcke, 'SRILM - An Extensible Language Modeling Toolkit', ICSLP 2002. Software available at http://www.speech.sri.com/projects/srilm [23] Academia Sinica, Part-of-Speech Tagger, http://ckipsvr.iis.sinica.edu.tw/ [24] Chao-Yu Huang, Lin-Shan Lee, “Language Model Adaptation for Mandarin-English Code-Mixed Lectures Using Word Classes and Random Forests”, NTU Master Thesis 2011 [25] The Snack Sound Toolkit, http:// http://www.speech.kth.se/snack/ [26] Hsin-Yi Lin, Janice Fon, “The Role of Pitch Reset in Perception at Discourse Boundaries”, ICPhs 2011 [27] C.-F. Yeh, C.-Y. Huang, L.-C. Sun, L.-S. Lee, 'An Integrated Framework for Transcribing Mandarin-English Code-Mixed Lectures with Improved Acoustic and Language Modeling', ISCSLP 2010 | |
| dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/63278 | - |
| dc.description.abstract | 語音處理的技術日新月異，從語音辨識率的提升至語意的理解分析，都被廣泛的研究及應用。語句分段可分為四個階段，首先用能量等基本特徵做粗分段，再將每一粗分段的語音訊號辨識成詞串，接著在詞與詞之間找出適當的語句邊界，最後再將重新分段的語句進行再辨識，以提升辨識率。本論文著重在第三階段的尋找適當斷點。
語音大致上可分為朗讀式語音和自發性語音兩大類，後者會因為語者的思路和語者習慣，說話節奏、韻律、及用詞上與前者不同。而本論文的研究語料為課程語音，屬於自發性語音，實驗中使用支撐向量機做為學習演算法訓練模型，並藉此比較不同的語彙及韻律特徵對語句分段的效用。 實驗分別做在人工標記轉寫和語音辨識後的結果上。語彙特徵包含語言模型特徵、詞性標記特徵、關鍵詞彙特徵，研究顯示語彙特徵彼此有加成性，其中又以關鍵詞彙特徵對結果的提升最有幫助，但由於實驗語料為單一語者，在多語者情況下，其效用尚待實驗；韻律特徵的部分使用了兩組不同的特徵，一組原用於音調辨識；另一組原用於偵測自發性語音中不流利處，而研究顯示後者較為有效。整體而言，韻律特徵的效果遠勝語彙特徵，但兩者具有一定程度的加成性。此外由於自發性語音中，語句邊界附近用詞特性的關係，人工標記轉寫和語音辨識結果的實驗數據和趨勢相近，說明辨識率對自發性語音的語句分段並無太大影響。 | zh_TW |
| dc.description.provenance | Made available in DSpace on 2021-06-16T16:32:18Z (GMT). No. of bitstreams: 1 ntu-101-R98922033-1.pdf: 637081 bytes, checksum: ae2368f6f74e62e4e7cb574209612ad7 (MD5) Previous issue date: 2012 | en |
| dc.description.tableofcontents | 目錄
中文摘要 i 目錄 ii 表目錄 iv 圖目錄 v Chapter1 導論 1 1.1 研究動機 1 1.2 相關研究 2 1.3 研究方法 4 1.4 章節安排 4 Chapter2 背景知識 6 2.1 語句分段簡介 6 2.2 機器學習 7 2.3 支撐向量機 9 2.4 N連語言模型 13 2.5 語句分段評估機制 15 2.6 本章總結 17 Chapter3 研究方法 18 3.1 語料介紹 18 3.1.1 實驗語料 18 3.1.2 語言模型訓練語料 19 3.2 系統架構與流程 19 3.3 語彙特徵 21 3.3.1 語言模型特徵 21 3.3.2 詞性標記特徵 22 3.3.3 關鍵詞彙特徵 23 3.4 韻律特徵 24 3.4.1 訊號維度 26 3.4.2 用於音調辨識 27 3.4.3 用於偵測自發性語音中不流利處 28 3.5 本章總結 30 Chapter4 實驗與結果 31 4.1 實驗設計 31 4.2 人工標記轉寫實驗結果 33 4.3 大字彙語音辨識實驗結果 41 4.4 本章總結 45 Chapter5 結論與展望 46 5.1 結論 46 5.2 未來展望 47 參考文獻 48 | |
| dc.language.iso | zh-TW | |
| dc.subject | 語彙特徵 | zh_TW |
| dc.subject | 語句分段 | zh_TW |
| dc.subject | 自發性語音 | zh_TW |
| dc.subject | 機器學習 | zh_TW |
| dc.subject | 韻律特徵 | zh_TW |
| dc.subject | machine learning | en |
| dc.subject | prosodic feature | en |
| dc.subject | sentence segmentation | en |
| dc.subject | lexical feature | en |
| dc.subject | spontaneous speech | en |
| dc.title | 使用支撐向量機的自發性語音語句分段 | zh_TW |
| dc.title | Sentence Segmentation of Spontaneous Speech using Support Vector Machine | en |
| dc.type | Thesis | |
| dc.date.schoolyear | 101-1 | |
| dc.description.degree | 碩士 | |
| dc.contributor.oralexamcommittee | 林軒田(Hsuan-Tien Lin),蘇雅韻(Ya-Yunn Su) | |
| dc.subject.keyword | 語句分段,自發性語音,機器學習,語彙特徵,韻律特徵, | zh_TW |
| dc.subject.keyword | sentence segmentation,spontaneous speech,machine learning,lexical feature,prosodic feature, | en |
| dc.relation.page | 50 | |
| dc.rights.note | 有償授權 | |
| dc.date.accepted | 2012-12-05 | |
| dc.contributor.author-college | 電機資訊學院 | zh_TW |
| dc.contributor.author-dept | 資訊工程學研究所 | zh_TW |
| 顯示於系所單位： | 資訊工程學系 | |
文件中的檔案：
| 檔案 | 大小 | 格式 | |
|---|---|---|---|
| ntu-101-1.pdf 未授權公開取用 | 622.15 kB | Adobe PDF |
系統中的文件，除了特別指名其著作權條款之外，均受到著作權保護，並且保留所有的權利。