簡易檢索 / 詳目顯示

研究生: 沈敬濠
Shen, Jing-Hao
論文名稱: 基於使用者音樂需求之音樂歌單推薦
Song List Recommendation based on User Music Needs
指導教授: 盧文祥
Lu, Wen-Hsiang
學位類別: 碩士
Master
系所名稱: 電機資訊學院 - 資訊工程學系
Department of Computer Science and Information Engineering
論文出版年: 2019
畢業學年度: 107
語文別: 英文
論文頁數: 61
中文關鍵詞: 音樂推薦歌詞分析歌曲列表使用者音樂需求歌曲結構人格特質情境分類歌詞聊天機器人支持向量機邏輯斯回歸
外文關鍵詞: Music Recommendation, Lyrics Analysis, Song List, User Music Needs, Song Structure, Personality Trait, Scene Classification, Lyrics Chatbot, Support Vector Machine, Logistic Regression
相關次數: 點閱:129下載:11
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報
  • 音樂一直是人們生活中不可缺少的元素,而音樂無所不在,隨時隨地都可以看見人們戴著耳機,聽著喜歡的音樂來度過自己的日常,伴隨著數位音樂串流的發達,使用者想要取得音樂也越來越方便,而華語流行音樂更是最受華人歡迎的歌曲主流,人們常常藉由聽歌來抒發心情,有許多不同的情境可以在華語音樂中所探索到,然而,目前許多知名的音樂串流公司紛紛推出不同的推薦歌單,例如音樂曲風或最新流行歌曲,但是可以供給使用者的選擇仍然不足。
    為了找出更多元的情境主題,我們回歸到聽音樂的需求,提出以使用者音樂需求作為出發點,欲建立不同的主題歌單,而什麼類型的主題會是使用者想要且符合他們的需求?我們歸納出三大面向,分別為Emotion Theme , Scene Theme以及Character and Personality Theme以便建立後續音樂歌單主題的建立的基準,為了有效符合使用者需求推薦歌單,我們建構Lyrics Chatbot,推薦歌曲的同時,找出符合使用者心境的歌詞,來吸引使用者作選擇。我們透過實驗證明,在分析歌詞結構中,以只考慮副歌為主的模型去做推薦時,在推薦首要的歌曲會表現的比只考慮主歌段落或者全部歌詞還要來的好。

    Music has always been an indispensable element in people's lives, and music is everywhere. You can see people wearing headphones and listening to your favorite music anytime, anywhere. With the development of digital music, users want it. It is more and more convenient to obtain music, and Chinese pop music is the most popular song among Chinese. People often express their feelings by listening to songs. There are many different situations that can be explored in Chinese music. However, many of them are currently Well-known music streaming companies have launched different recommended song lists, such as music styles or the latest popular songs, but the choices available to users are still insufficient.
    In order to find out more about the situational theme, we return to the need to listen to music, propose to use the user's music needs as a starting point, to create different theme songs, and what type of theme will be what the user wants? We summed up the three major aspects, respectively, "Emotion Theme", "Scene Theme" and "Character and Personality Trait Theme". In order to establish the theme of the follow-up music song list benchmarks and effectively match the user's needs to recommend songs, we built Lyrics Chatbot, while recommending songs, find the lyrics that match the user’s mood to attract users to make choices. Through experiments, we prove that in the analysis of the lyrics structure, when we make song recommendations based on models that only consider chorus, it is better to recommend the first song than to consider only the verse or whole lyrics.

    摘要 III Abstract V 致謝 VII Tables of Contents VIII List of Tables X List of Figures XII Chapter 1 Introduction 1 1.1 Background 1 1.2 Motivation 3 1.3 Method 3 1.4 Contribution 5 1.5 Organization of this Dissertation 5 Chapter 2 Related Work 6 2.1 Studies on Exploring Song Lyrics Structures 6 2.2 Studies on User Music Needs 7 Chapter 3 Method 8 3.1 System Framework 8 3.2 Preliminaries 11 3.2.1 CKIP Word Segmentation System 11 3.2.2 Extended-HowNet 11 3.2.3 Datasets 13 3.3 Song Structure Analysis 15 3.4 Emotion Theme Extraction 19 3.5 Scene Theme Extraction 21 3.5.1 Scene from Observation 22 3.5.2 Scene from User Music Needs Question 28 3.6 Character and Personality Theme Extraction 30 3.6.1 Person Pronoun Analysis 30 3.6.2 Personality Traits 32 3.7 Scene Entity Generation Model 34 3.8 Song Recommendation based on User Music Needs 38 3.9 Lyrics Chatbot 40 Chapter 4 Experiments 41 4.1 Dataset 41 4.2 Experiment of Love and Friend Scene Classification 42 4.2.1 Dataset for Love and Friend Scene Classification 42 4.2.2 Evaluation Metrics 43 4.2.3 Experiment Result 43 4.3 Experiment of Scene Entity Threshold 45 4.3.1 Dataset for Evaluate Scene Entity Threshold 45 4.3.2 Evaluation Metrics 46 4.3.3 Experiment Result 47 4.4 Evaluation on Song Recommendation of Verse & Chorus 47 4.4.1 Dataset for Evaluation on Song Recommendation of Verse & Chorus 48 4.4.2 Evaluation Metrics 49 4.4.3 Experiment Result 51 Chapter 5 Conclusions 57 Reference 58

    [1] Schäfer T, Sedlmeier P, Städtler C, Huron D. The psychological functions of music listening. Front Psychol. 2013
    [2] 蔡振家(Chen-Gia Tsai), 陳容姍(Rong-Shan Chen) &余思霈(Szu-Pei Yu).解析「主歌-副歌」形式:抒情慢歌的基模轉換與音樂酬賞 Analyzing the Verse-Chorus Form: Schema Shifts and Musical Rewards in Lyrical-Slow Songs,P239 - 286
    [3] WHICH SONG STRUCTURE SHOULD YOU CHOOSE? CareersInMusic(2019)
    https://www.careersinmusic.com/song-structure/
    [4] 歌曲結構分析:Go Chic-DJ Sonia, https://ysolife.com/compose/
    [5] J. P. G. Mahedero et al., “Natural language processing of lyrics,” in Proc. ACM MM, pp. 475–478, 2005.
    [6] N. C. Maddage et al., “Content-based music structure analysis with applications to music semantics understanding”, in Proc. ACM MM, pp. 112–119, 2004
    [7] M. Casey et al., “Content-based music information retrieval: current directions and future challenges,” in Proceedings of the IEEE, vol. 96, no. 4, pp. 668–696, 2008.
    [8] J. Paulus and A. Klapuri, “Music structure analysis by finding repeated parts,” in Proc. ACM Audio and Music Computing for Multimedia Workshop, pp. 59–68, 2006.
    [9] Cai, R., Zhang, C., Wang, C., Zhang, L., & Ma, W. Y. (2007, September). Musicsense: contextual music recommendation using emotional allocation modeling. In Proceedings of the 15th ACM international conference on Multimedia (pp. 553-556). ACM.
    [10] Logan, B., Kositsky, A., & Moreno, P. (2004, June). Semantic analysis of song lyrics. In Multimedia and Expo, 2004. ICME'04. 2004 IEEE International Conference on (Vol. 2, pp. 827-830). IEEE.
    [11] Fell, M., & Sporleder, C. (2014). Lyrics-based Analysis and Classification of Music. In COLING (Vol. 2014, pp. 620-631).
    [12] M. Haro, A. Xamb´o, F. Fuhrmann, D. Bogdanov, E. G´omez, and P. Herrera. The musical avatar – a visualization of musical preferences by means of audio
    content description. In Audio Mostly (AM ’10), Pitea, Sweden, 2010. ACM.
    [13] M. Levy and M. Sandler. Music information retrieval using social tags and audio. IEEE Transactions on Multimedia, 11(3):383–395, 2009.
    [14] Q. Li, S. H. Myaeng, and B. M. Kim. A probabilistic music recommender considering user opinions and audio features. Information Processing & Management, 43(2):473–487, Mar. 2007.
    [15] E. Pampalk. Computational models of music similarity and their application in music information retrieval. PhD thesis, Vienna University of Technology, Mar. 2006.
    [16] M. Slaney and W. White. Similarity based on rating data. In International Symposium on Music Information Retrieval (ISMIR’07), 2007.
    [17] Brattico E, Alluri V, Bogert B, Jacobsen T, Vartiainen N, Nieminen S and Tervaniemi M (2011) A functional MRI study of happy and sad emotions in music with and without lyrics.
    [18] C. Laurier and J. Grivolla. Multimodal Music Mood Classification Using Audio and Lyrics. In Int. Conf. Machine Learning and Applications, pages 1–6, 2008.
    [19] B. Schuller, J. Dorfner, and G. Rigoll. Determination of Nonprototypical Valence and Arousal in Popular Music: Features and Performances. EURASIP Journal on Audio, Speech, and Music Processing, 2010:1–19, 2010.
    [20] Menno van Zaanen and Pieter Kanters. AUTOMATIC MOOD CLASSIFICATION USING TF*IDF BASED ON LYRICS. In J. Stephen Downie and Remco C. Veltkamp, editors, 11th International Society for Music Information and Retrieval Conference, August 2010.
    [21] X. Hu, J. S. Downie, and A. F. Ehmann, “Lyric text mining in music mood classification,” in Proc. of the Intl. Society for Music Information Retrieval Conf., Kobe, Japan, 2009.
    [22] CKIP, http://ckipsvr.iis.sinica.edu.tw/
    [23] Extended-HowNet, http://ehownet.iis.sinica.edu.tw/index.php
    [24] CKIP Group (2013 reversion), Lexical Semantic Representation and Semantic Composition- An Introduction to E- How Net, CKIP Technical Report.
    [25] Huang, S. L., Chung, Y. S., & Chen, K. J. (2008). E-HowNet- an Expansion of HowNet. In Proceedings of 1st National HowNet Workshop, Beijing, China.
    [26] Sahlgren, M., & Cöster, R. (2004, August). Using bag-of-concepts to improve the performance of support vector machines in text categorization. In Proceedings of the 20th international conference on Computational Linguistics (p. 487). Association for Computational Linguistics.
    [27] Nigam, Kamal et al. “Text Classification from Labeled and Unlabeled Documents using EM.” Machine Learning 39 (2000): 103-134.
    [28] Ding, Chris H. Q. and Inna Dubchak. “Multi-class protein fold recognition using support vector machines and neural networks.” Bioinformatics 17 4 (2001): 349-58 .
    [29] Wu, Ting-Fan et al. “Probability Estimates for Multi-class Classification by Pairwise Coupling.” J. Mach. Learn. Res. 5 (2003): 975-1005.
    [30] Lee, Jin Ha and J. Stephen Downie. “Survey Of Music Information Needs, Uses, And Seeking Behaviours: Preliminary Findings.” ISMIR (2004).

    下載圖示 校內:2021-08-27公開
    校外:2021-08-27公開
    QR CODE