Title page for etd-0522119-231009
 論文名稱Title 基於機器學習之支持向量機於演唱會產業變數分析An analysis of features of the total revenue of the concerts based on Machine Learning (Support Vector Machine) 系所名稱Department 企業管理學系Department of Business Management 畢業學年期Year, semester 107 學年度 第 2 學期The spring semester of Academic Year 107 語文別Language 中文Chinese 學位類別Degree 碩士Master 頁數Number of pages 58 研究生Author 林育慈Yu-Tzu Lin 指導教授Advisor 召集委員Convenor 口試委員Advisory Committee 口試日期Date of Exam 2019-06-19 繳交日期Date of Submission 2019-06-22 關鍵字Keywords 音樂產業、演唱會、機器學習、支持向量機、套索演算法Lasso Regression, Support Vector Machine, Machine Learning, Music industry, Concert 統計Statistics 本論文已被瀏覽 5812 次，被下載 44 次The thesis/dissertation has been browsed 5812 times, has been downloaded 44 times.
 中文摘要 音樂演唱會的總收益，向來是音樂產業內部進行票價訂定時的標準之一；假設某策展團體預計邀請一位從未造訪該國之音樂人進行演出時，許多人都認為Youtube歌曲瀏覽量能夠幫助他們預測總收益。此外，其網路聲量、演出地點、是否發行新歌等變數出發，對於策展人來說，也是考量市場是否有足夠需求之重要特徵。本研究據以機器學習之支持向量機演算法，配合模糊理論與統計分類方法，將所蒐集到之二次資料進行變數分群，歸納並預測出不同總收益之音樂人的特徵變量為何。透過蒐集兩年期之Pollstar資料，並建立在演唱會產業之業界觀點（覺醒音樂祭 Wake Up Festival / 傲頭厝展演空間 Our Town Love House），來發想此研究內容。研究結果顯示，據本研究所蒐集之變量特徵來將音樂人族群進行分群，若是音樂人能透過Youtube歌曲瀏覽量及其他特徵變數分類，得知該場演唱會的總收益分群將落在哪個區間，將可符合在地市場之趨勢與需求，便於分析與預測營收，並從加以訂價，從中創造收益。 Abstract Concert tickets revenue has always been a component of the pricing mechanism. When selecting among foreign performers, most of the concert promoters consider the views on Youtube to be helpful for building pricing model. Furthermore, to promoters, other factors including internet word of mouth, exhibition location or new-released albums are also crucial attributes regarding the demand in music industry. That’s why this research, based on the SVM model, intend to manifest what are the features that are able to predict the potential revenue. The study is developed through the combination of 2-year data on Pollstar and the practical point of view from “Wake Up Festival” and “Our Town Love House.” The indications is: the collected features could cluster the performers; phrased otherwise, applying the views on Youtube and other related features, promoters could better conduct the price mechanism and create value since the local market trend and demand are analyzed and well-understood.
 目次 Table of Contents 目錄論文審定書............................................................................................................................................................i誌謝....................................................................................................................................................................... ii摘要.......................................................................................................................................................................iiiAbstract................................................................................................................................................................iv目錄.........................................................................................................................................................................v圖次........................................................................................................................................................................vi表次.................................................................................................................................................................... viii 第一章 緒論.................................................................................................11.1 研究背景................................................................................................11.2 研究動機與缺口.....................................................................................21.3 研究問題................................................................................................3第二章 文獻回顧..........................................................................................42.1 音樂演唱會產業.....................................................................................42.2 篩選特徵變數演算法.............................................................................62.3 支持向量機............................................................................................7第三章 研究方法........................................................................................153.1 研究模型介紹......................................................................................153.2 資料來源..............................................................................................173.2.1 主要資料來源：Pollstar、Kworb....................................................173.2.2 變數來源...........................................................................................193.2.3 資料清洗...........................................................................................223.3 變數介紹..............................................................................................233.4 篩選變數演算法..................................................................................253.5 LASSO REGRESSION.........................................................................263.5.1 L1正則化.........................................................................................263.6支持向量機.........................................................................................28第四章 研究結果......................................................................................294.1 資料敘述統計....................................................................................294.1.1變數之資料敘述統計......................................................................294.2 篩選特徵變數演算法........................................................................364.2.1 Lasso Regression.........................................................................364.4支持向量機結果................................................................................39第五章 結論............................................................................................425.1 研究結論..........................................................................................425.2 研究限制..........................................................................................43參考文獻................................................................................................44圖次圖2-1 Lasso Regression 之L1 正則化.................................................................................................. 7 圖2-2 支持向量機超平面示意圖.............................................................................................................9 圖2-3 支持向量機最大間隔超平面示意圖...........................................................................................10圖3-1 演唱會特徵分析架構圖...................................................................................................................16圖3-2 POLLSTAR 官網操作介面..........................................................................................................18圖3-3 KWORB.NET 官網操作介面......................................................................................................19圖3-4 LIVE BOXOFFICE DATABASE ..............................................................................................20圖3-5 GLOBAL CONCERT PULSE .......................................................................................................22圖3-6 挑選最佳懲罰值(LAMBDA) ..........................................................................................................27 圖4-1 訓練集總收益之資料分佈(取對數) ...........................................................................................34 圖4-2 測試集總收益之資料分佈(取對數) ........................................................................................... 34 圖4-3 訓練集之總收益QQ圖....................................................................................................................35圖4-4 測試集之總收益QQ圖 .................................................................................................................35表次表3-1 POLLSTAR BOXOFFICE提供之變數....................................................................................21表3-2 變數介紹...............................................................................................................................................23 表3-3 變數因素向量化.................................................................................................................................26 表4-1 場次及日期之敘述統計表 .............................................................................................................30 表4-2 虛擬變數之敘述統計表...................................................................................................................32 表4-3 分群資料之敘述統計表...................................................................................................................33 表4-4 LASSO REGRESSION 之RMSE...............................................................................................36 表4-5 LASSO REGRESSION 篩選變數...............................................................................................37 表4-6 LASSO REGRESSION 挑選之子集合.....................................................................................38 表4-7 訓練集之混淆矩陣.............................................................................................................................40 表4-8 測試集之混淆矩陣.............................................................................................................................40表4-9 訓練集與測試集之分群預測準確度...........................................................................................41表4-10 精確率、召回率、F1 Score.........................................................................................................41表5-1 高低收益分群之特徵變數整理....................................................................................................42
 參考文獻 References 1. 英文文獻Vapnik V. (1995). The Nature of Statistical Learning Theory[M]. New York: Springer. A. Belloni, V. Chernozhukov, and C. Hansen (2010). Lasso methods for gaussian instrumental variables models. MIT Department of Economics Working Paper, No.11-14Mehmet Caner. (2010). The Adaptive Lasso Method for Instrumental Variable Selection. Department of Economics, North Carolina State University.DA. Ackerberg. (2003). Advertising, learning, and consumer choice in experience goodmarkets: an empirical examination.International Economic Review, 44(3), 1007-1040. Alan B. Krueger. (2005). The Economics of Real Superstars: The Market for Rock Concerts in the Material World. Journal of Labor Economics, Vol. 23, No. 1, pp. 1-30. Alan B. Krueger & Marie Connolly (2005). Rockonomics:The Economics of popular music. National Bureau of economic research, JEL No. Z1, L82, O34. Caves, R. E. (2000). Creative industries: Contracts between art and commerce. Harvard University Press. Court, Andrew T. (1939). Hedonic Price Indexes with Automotive Examples. New York General Motors Dewenter (2012). On File Sharing With Indirect Network Effects Between Concert Ticket Sales and Music Recordings. Journal of Media Economics. Frank ,R & Cook, K. S. (1995). The winner-take-all society.Hamlen, William A, Jr (1991). "Superstardom in Popular Music: Empirical Evidence, The Review of Economics and Statistics, MIT Press, vol. 73(4), pages 729-733, November.Holt F. (2010). The economy of live music in the digital age, European Journal of Cultural Studies.Kelvin J. Lancaster (1965). A New Approach to Consumer Theory. Journal of Political Economy, Vol. 74, No. 2 (Apr., 1966), pp. 132-157 Meier, G & Rivkin, J.W. (2000). BMG Entertainment,Harvard Business School Case, 701-003. Rosen, S. (1981). The economics of superstars, The American economic review, 71(5), 845-858. Rosen, S. (1974). Hedonic prices and implicit markets product differentiation in pure competition. Journal of Political Economy. Simon. F. (2001). The popular music industry, The Cambridge companion to pop and rock. Cambridge University Press. Simon. F. (2010). Analysing live music in the UK:findings one year into a three-year research project. Journal of the international association for the study of popular music. Staiger & Stock (1997). Instrumental Variables Regression with Weak Instruments. Econometrica. Throsby. (1994). The Production and Consumption of the Arts: A View of Cultural Economics. Journal of Economic Literature. William W. Fisher III (2004). Promises to keep: technology, law and the future of entertainment. Stanford University Press. 2. 中文文獻梁定澎. (2018). 人文社會大數據研究. 國立中山大學電子商務研究中心.顧謙. (2018). Python深度學習實戰. 機械工業出版社.Sebastian Raschka. (2018). Python機器學習. 博碩文化股份有限公司.陳仁和. (2017). 初探機器學習|使用Python. 碁峰資訊股份有限公司.陳承昌. (2017). 支持向量機應用於水稻田辨識之研究. 航測及遙測學刊. 第12卷. 第3期. 頁225-240.財團法人臺灣唱片出版事業基金會. (2017). 台灣唱片發展現況.Michael Bowles. (2016). 機器學習|使用Python進行預測分析的基本技術. 碁峰資訊股份有限公司.范淼. (2016). Python 機器學習及實踐. 清華大學出版社.李明璁. (2015). 樂進未來. 大塊文化出版股份有限公司.財團法人臺灣唱片出版事業基金會. (2013). 台灣唱片發展現況. 李航. (2012). 統計學習方法. 清華大學出版社. 頁95王偉霖,楊珮琪. (2011). 文化創意產業與中國大陸法制─以音樂產業為中心. 元照出版公司. 林富美. (2009). 台灣音樂歹年冬覓出路. 行政院新聞局. 頁 217.黃福瑞. (2003). 從矽谷到好萊塢─企業的 e 化與 E 化. 中國生產力中心. 頁 113.3. 碩士論文米麗曼. (2018). 基於機器學習與AH溢價指數漲跌的預測.國立台灣大學資訊工程學研究所碩士論文.吳胤霆. (2018). 演唱會策展人因素對演唱會市場需求影響.國立中山大學企業管理人研究所碩士論文.莊皓宇. (2018). 應用支持向量機於動漫圖像分類. 國立台灣大學農業學研究所碩士論文.廖君琳. (2018). 演唱會購票意願之影響因素：以計劃行為理論之觀點探討. 國立成功大學企業管理學研究所碩士論文.鍾鎮宇. (2018). 是看直播呢還是聽現場？閱聽人對演唱會臨場感接受研究. 世新大學傳播管理學研究所碩士論文.楊竣翔. (2017). 使用支持向量機技術之人臉影像辨識. 國立勤益科技大學電子工程學研究所碩士論文.蘇彥廷. (2016). 支持向量機模型在台灣加權股價指數趨勢之預測. 國立中山大學財務管理學研究所碩士論文.姜佩君. (2016). How Do I Live : 臺灣流行音樂消費「現場」中的儀式參與與認同過程.國立台灣大學社會學研究所碩士論文.蘇鈺娟. (2016). 流行音樂演唱會科技應用之發展策略研究.國立臺北藝術大學藝術行政與管理學研究所碩士論文.黃智煌. (2016). 運用函數型資料分群方法與類神經網路於演唱會票卷銷售預測之研究.國立臺灣科技大學工業管理學研究所碩士論文.鄭敦維. (2012). 一個基植於遺傳演算法與模糊理論最佳化之支援向量機選股模型.國立高雄大學資訊工程研究所碩士論文.蘇紀瑋. (2012). 現場演唱會之最適訂價研究.國立台灣大學企業管理碩士專班碩士論文.羅方翰. (2010). 網路影音平台對台灣流行音樂生產與行銷的影響之初探. 國立政治大學傳播學院碩士在職專班碩士論文.張容瑛. (2008). 華文流行音樂區域與都市形構及其治理. 國立臺北大學都市計劃研究所碩士論文.高珮芳. (2004). 線上音樂商店系統架構之競爭分析.國立政治大學智慧財產研究所碩士論文.
 電子全文 Fulltext 本電子全文僅授權使用者為學術研究之目的，進行個人非營利性質之檢索、閱讀、列印。請遵守中華民國著作權法之相關規定，切勿任意重製、散佈、改作、轉貼、播送，以免觸法。論文使用權限 Thesis access permission：自定論文開放時間 user define開放時間 Available：校內 Campus： 已公開 available 校外 Off-campus： 已公開 available 紙本論文 Printed copies 紙本論文的公開資訊在102學年度以後相對較為完整。如果需要查詢101學年度以前的紙本論文公開資訊，請聯繫圖資處紙本論文服務櫃台。如有不便之處敬請見諒。開放時間 available 已公開 available
 QR Code