透過您的圖書館登入
IP:3.129.39.55
  • 學位論文

確認影響直播觀看次數關鍵因素之研究-以Twitch.tv為例

Identifying the Key Factors of Views in Live Streaming: An Example of Twitch.tv

指導教授 : 陳隆昇
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


隨著資訊的日新月異,直播(Live Streaming)服務在全球迅速發展,又因技術限制的減少、直播平台逐漸興盛,使得人人都能成為直播者。在這樣的情況下,具有大量的使用者創造內容(User-generated content, UGC)的直播環境,由於內容豐富多彩,逐漸成為大多數人的休閒活動之一,因此,在觀看其他人在直播平台上玩遊戲的互動直播方式也越來越受歡迎。在過去的文獻中,大多數對於直播的研究主要集中在預測直播期間觀眾的數量或找出受歡迎的直播者,近年來則多半在探討在使用者的行為動機。如:直播中的禮物贈送行為。然而,從現有文獻中,相對較少研究集中在討論影響使用者觀看行為的聊天或評論。因此,本研究試圖運用直播聊天室中的評論進行實驗分析,透過最小絕對壓縮挑選機制(LASSO)、支持向量機之遞迴式特徵消除(SVM-RFE)與卡方檢定(Chi-square test)等特徵選取方法,來找出影響直播觀看次數之重要詞彙,接著,運用K-means分群和潛在語意分析(LSA)之奇異值分解(SVD)進行概念化分類,以找出影響觀看直播次數之重要因素。分析結果希望能夠做為直播平台業者及直播頻道的參考。

並列摘要


With the update of information, live streaming has developed rapidly in the world, coupled with the reduction of technical restrictions, so that everyone can become a streamer. Due to the rich and colorful content, it has become one of the leisure activities of most people. The method of watching other people playing games on the live streaming platform is also becoming more and more popular. In the past academic literatures, most of the researches mainly focused on predicting the number of viewers during live streaming or finding popular streamer. Such as: the gift giving behavior in the live streaming. However, from the existing literature, relatively few studies have focused on chats or comments that affect users' viewing behavior. Therefore, this study attempts to use the comments in the live chat text for experimental analysis, through the least absolute shrinkage and selection operator (LASSO), support vector machine-recursive feature elimination (SVM-RFE) and chi-square test to find the important vocabularies that affects the number of live streaming views, and then use K-means and singular value decomposition (SVD) of latent semantic analysis (LSA) for conceptual classification to find important factors that affect the views of live streaming. The analysis results hope to be a reference for the live streaming platform and streamers.

參考文獻


[1] 林孟儒(2017),確認影響手機遊戲App內消費之關鍵因素,碩士論文,朝陽科技大學資訊管理系,臺中。
[2] 曾恕慈(2020),旅遊業顧客不再訪重要因素分析之研究 - 以TripAdvisor飯店評論為例,碩士論文,朝陽科技大學資訊管理系,臺中。
[3] 鄭紹偉(2009),集成分類器結合特徵選取與多字詞判斷疾病分類碼,碩士論文,國立雲林科技大學資訊管理系,雲林。
[4] 蕭伃婷(2018),應用文字探勘技術於旅館服務特性品質診斷,碩士論文,國立臺北大學企業管理學系,臺北。
[5] 簡婉婷(2018),影響離線商務模式信任因素確認之研究 - 以OpenTable為例,碩士論文,朝陽科技大學資訊管理系,臺中。

延伸閱讀