影劇歌曲活動事件與歌手網路聲量關係之擷取與分析( I );Extraction and Analysis of Artist-Relevant Events and Share of Voice from Social Network( I )

NCUIR > College of Electrical Engineering & Computer Science > Department of Computer Science and information Engineering > Research Project > Item 987654321/82327

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/82327

Title:	影劇歌曲活動事件與歌手網路聲量關係之擷取與分析( I );Extraction and Analysis of Artist-Relevant Events and Share of Voice from Social Network( I )
Authors:	張嘉惠;莊秀敏
Contributors:	國立中央大學資訊工程學系
Keywords:	命名實體辨識;Named entity recognition
Date:	2020-01-13
Issue Date:	2020-01-13 14:40:37 (UTC+8)
Publisher:	科技部
Abstract:	對於唱片業、娛樂事業而言，發掘有潛力的新人、了解其市場的關注度、規劃其生涯發展的規劃都是一連串的決策過程。過去這些新人的培植過程有賴經紀人的經驗，若是能配合充足與正確的資訊，會讓星探工作獲得更有力的支持，對於KKTV以及KKBOX等整合影音、音樂串流平台而言，有其重要性。然而相對於棒球的每場比賽均有相關資料的記錄，過去新人的發片、簽唱會、代言等所造成的市場反應度資訊並不容易取得。不過由於社群網路的蓬勃發展，這些相關事件、以及市場反應度，我們可以透過社群網路意見探勘來掌握市場動向，不僅可以幫助其經紀公司記錄其下藝人發展，也可以做為產業行銷與決策的考量。本計畫預計包含兩大主題：其一是市場反應度分析、其次是藝人活動事件的擷取。市場反應度分析的關鍵技術又分為：關注對象的辨識、以及情緒分類。換言之，市場反應度分析不僅僅是命名實體辨識，還要辨識出語句中的意見對象，以精準評估使用者對於藝人或對其作品的正負面情緒。由於戲劇、歌曲、或歌手網路聲量的變動往往伴隨著某個事件的或活動的發生，例如簽唱會、代言、發片、公演等，因此本計畫的第二年即著重在社群網路上的活動、事件的擷取，並且鏈結每個活動、事件、或是作品至相對的Artist。 ;For the record and entertainment industries, fostering potential artist, creating artworks to attract market attention is a series of decision-making processes. In the past, the excavation and cultivation process of these singers or actors depended on the experience of the brokers. If we have sufficient and correct information of the artists’ market attention, it will be a powerful support for the scouting work for video and music streaming platforms like KKTV and KKBOX. Thus, identifying market reaction from social network is an important task for the entertainment industry.To achieve this goal, we define two major tasks to identify market reaction from social network: one is market attention analysis, the other is discovering the cause of significant change in market attention. For market attention analysis, the tasks can be further divided into sentiment classification of opinion statements and identification of opinion target such as artist name and song/drama/movie name. In other words, the market attention analysis is not only to recognize entity names, but also to ensure whether it is the target in the opinion statement. Meanwhile, the changes in the market attention are often accompanied by artist related events or activities, such as autograph session, concert, performance in a film, endorsement, etc. Thus, the second year of the project focuses on extraction of artist activity and event as well as the relation to the artist.
Relation:	財團法人國家實驗研究院科技政策研究與資訊中心
Appears in Collections:	[Department of Computer Science and information Engineering] Research Project

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	207	View/Open

社群 sharing

Loading...