The Evolution of Taiwan Mandopop Music Lyric Mood by Sentiment Analysis

• Main Authors: LIAO, WEI-FAN, 廖偉帆
• Other Authors: Maria R. Lee
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/56750719537348208740

碩士 === 實踐大學 === 資訊科技與管理學系碩士班 === 104 === With the advent of the era of mobile network, music-streaming services, such as KKBOX, Spotify, and Apple Music, have changed music audience’s listening music behaviors. The aim of this study is to explore KKBOX Mandopop music’s trend development, music lyrics emotion preferences and trend analysis. We crawled daily top one hundred Mandopop music in KKBOX music steaming between the year of 2005 - 2015. More than 300,000 leaderboard data and 5,043 lyrics data have been collected. We applied SVM (Support Vector Machine) algorithm for music emotion classification, lyric-based song sentiment analyzing for Chinese text-mining, and visualization techniques for the trend analysis. The research results show that (i) most of Mandopop music’s artists come from Taiwan for the past 10 years; (ii) The songs in dramas and movies’ have major influences to music popularity; (iii) Most songs in leaderboard have positive emotion. The findings of this study may support the music-streaming services to better understand their audience, and provide a guide reference for the traditional music industry.