資料探勘技術在電影推薦上的應用研究-以F線上影音平台為例

NCU Institutional Repository > 管理學院 > 資訊管理學系碩士在職專班 > 博碩士論文 > Item 987654321/77569

請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/77569

題名:	資料探勘技術在電影推薦上的應用研究-以F線上影音平台為例
作者:	蔡明仁;Tsai, Ming-Jen
貢獻者:	資訊管理學系在職專班
關鍵詞:	線上影音平台推薦系統;資料探勘;分群分析;關聯分析
日期:	2018-06-29
上傳時間:	2018-08-31 14:48:42 (UTC+8)
出版者:	國立中央大學
摘要:	隨著互聯網技術的日益成熟以及娛樂文化的發展，人們越來越依賴通過數位影音平台觀看電影、戲劇。在資訊爆炸時代，信息量急劇上升，影音平台的資訊也日益豐富，如何在浩如煙海資料中找到人們有興趣觀看的電影成為一個需要被解決的課題。電影作為日常主要精神娛樂項目之一也存在資訊超載的難題。本研究為達到個性化推薦的目的，每日對線上影音平台上的用戶基本資料、電影基本資料及用戶對電影的觀看行為，使用資料探勘技術對用戶進行分群分析，使同一個群集的用戶有最相似的電影觀看興趣，使不同群集的用戶有最大差異的電影觀看喜好，並在用戶進入平台時向其推薦本人未曾觀看且同一群集的用戶曾經觀看過次數排前10名的電影。將同群集用戶觀看的電影套用關聯分析中的購物籃分析概念，以同一用戶每二個月觀看電影的行為當做一次購物籃的交易來分析同群用戶通常看什麼電影，哪些電影經常會被連續觀看，以及看完某影片後會續看那些影片。本研究同時透過關聯法則對2018年1-2月份用戶觀看電影的行為進行分析，得出F線上影音大多數的用戶都喜歡看2011年由美國出版的恐怖片。關鍵詞：線上影音平台推薦系統、資料探勘、分群分析、關聯分析;With the increasing maturity of Internet technology and the development of entertainment culture, people are increasingly relying on watching movies and dramas through digital video platforms. In the era of information explosion, the amount of information has risen sharply, and the information on the video and audio platforms has become increasingly abundant. How to find movies that people are interested in watching in a vast array of materials has become a topic that needs to be addressed. As one of the major daily spiritual entertainment projects, movies also have the problem of information overload. In order to achieve the purpose of personalized recommendation, this study uses user data, platform basic data, and user′s viewing behavior on the online video platform, and uses data mining technology to analyze users in groups so that users in the same cluster have The most similar movie viewing interests allow the users of different clusters to have the most disparate movie viewing preferences and recommend to the user that the user of the same cluster has never watched the top 10 movies when he enters the platform. Applying the shopping basket analysis concept in association analysis to the movie watched by the cluster user, the same user viewing the movie every two months as a shopping basket transaction to analyze what movies the same group of users usually see, and which movies are often consecutive Watch and watch those videos after watching them. This study also analyzed the behavior of users watching movies in January-February 2018 through association rules. The majority of users who found that audio and video on the F line like to watch horror films published in the United States in 2011. Keywords: online video platform recommendation system, data exploration, cluster analysis, association analysis
顯示於類別:	[資訊管理學系碩士在職專班 ] 博碩士論文

文件中的檔案:

檔案	描述	大小	格式	瀏覽次數
index.html		0Kb	HTML	231	檢視/開啟

在NCUIR中所有的資料項目都受到原著作權保護.

社群 sharing

資料載入中.....