強化使用者和電影評分關係，打造 User profile 之電影推薦系統

簡易檢索 / 詳目顯示

回結果列表

研究生：	馬曼容 Man-Jung Ma
論文名稱：	強化使用者和電影評分關係，打造 User profile 之電影推薦系統
指導教授：	陳彥良 Yen-Liang Chen
口試委員:
學位類別：	碩士 Master
系所名稱：	管理學院 - 資訊管理學系 Department of Information Management
論文出版年：	2020
畢業學年度：	108
語文別：	中文
論文頁數：	47
中文關鍵詞：	推薦系統、電影、個人化檔案、正面使用者檔案、負面使用者檔案
相關次數：	點閱：14 下載：0
分享至:	分享至facebook 分享至twitter

查詢本校圖書館目錄查詢臺灣博碩士論文知識加值系統勘誤回報

推薦系統廣泛應用於電影平台上，最廣泛的推薦方法是通過蒐集多個使用者所提供的觀看信息，以評分(Ranking-based)作為依據，與有相同興趣使用者進行比對，根據相似鄰居對項目給出的評分，計算出使用者之間的偏好相似性。最終，對該使用者尚未評分的項目，進行預測性的推薦，亦對使用者提供觀影上更多且更好的建議。

然而，儘管以大數據評分為根基，這樣以評分為基礎的電影推薦系統，卻無法細微觀察單一使用者的觀影喜好，甚至以擁有共同經驗之群體喜好作依據，卻不再以個人的觀看或評級紀錄，作為推薦的主要目的，更忽略電影資訊擁有的獨特性區別。

本文希望通過正面/負面使用者檔案的建立，進行多元指標的主題分析，增加一部電影的獨特性特徵，並納入過往被忽略的使用者低分評級紀錄，完成更客製且精準的個人化推薦方法。

Recommendation System is widely used on Movie over-the-top platforms. The most widely recommended method is to collect the viewing information provided by multiple users, and use the ratings as the basis to compare with users in the same interests. Based on the ratings given by similar neighbors to the project, the similarity of preferences between users is calculated. In the end, predictive recommendations are made to items that the user has not yet rated, and more and better suggestions are provided to the user.

However, despite the big data score as the foundation, this kind of rating-based recommendation system can not observe the single user's viewing preferences, even based on the group preferences with common experience, but no longer watched by individuals. Or ratings, as the main purpose of the recommendation, ignore the unique differences in film information.

This paper hopes to establish a multi-indicator theme analysis through the establishment of positive and negative user files, increase the unique characteristics of single movie, and incorporate the previously ignored users' low ratings to complete more customized and accurate personalized recommendation system.

中文摘要 ........................................................................................................................... I
英文摘要 ..........................................................................................................................II 
圖目錄 ............................................................................................................................ VI 表目錄 ........................................................................................................................... VII 一、緒論 .......................................................................................................................... 1
1 研究背景 ........................................................................................................... 1
2 研究動機 ........................................................................................................... 3
3 研究目的 ........................................................................................................... 4

二、文獻探討 .................................................................................................................. 7
1 以評分為基礎的方法(RATING-BASED)....................................................... 7
2 協同過濾(COLLABORATIVE FILTERING)....................................................... 7
2.1 基於使用者的協同過濾(User-based) .................................................... 8
2.2 基於產品的協同過濾(Item-based) ........................................................ 9
2.3 矩陣分解法(Matrix Factorization) ......................................................... 9
3 基於內容過濾(CONTENT-BASED FILTERING) .............................................11
4 混合性推薦 ..................................................................................................... 12
5 本研究方法介紹 ............................................................................................. 14

三、研究方法 ................................................................................................................ 16
1 協同過濾 ......................................................................................................... 16
2 潛在狄利克里分配(LATENT DIRICHLET ALLOCATION) ............................. 17
3 詞語頻率-逆向文件頻率(TF-IDF) ............................................................. 20
4 基於正面/負面使用者檔案方法 ................................................................. 22
5 KL-DIVERGENCE 和 JENSEN-SHANNON DIVERGENCE ...................................... 23

四、實驗設計 ................................................................................................................ 25
1 實驗資料集 ..................................................................................................... 25
2 預處理——比例組合 ..................................................................................... 26
3 實驗設計 ......................................................................................................... 26
3.1 nDCG 指標............................................................................................. 27 4.4 實驗結果......................................................................................................... 29
4.1 預處理 .................................................................................................... 29
4.2 正面/負面使用者檔案相似性 ............................................................ 30
4.3 nDCG 指標............................................................................................. 32

五、結論 ........................................................................................................................ 33
1 研究發現......................................................................................................... 33
2 研究限制和未來展望..................................................................................... 34
參考文獻 ........................................................................................................................ 35
                                

〔1〕 Symeonidis P. (2008) Content-based Dimensionality Reduction for Recommender Systems. Data Analysis, Machine Learning and Applications pp. 619-626: Springer.
〔2〕 Breese JS, Heckerman D, Kadie C (1998) Empirical analysis of predictive algorithms for collaborative filtering. Proc 40th Conf Uncertain AI 1:43–52
〔3〕 Sarwar B, Karypis G, Konstan J, Riedl J (2001) Item-based collaborative filtering recommendation algorithms. Proc 10th Int Conf WWW 1:285– 295.
〔4〕 Zhao, Changwei; Sun, Suhuan; Han, Linqian; Peng, Qinke (2016). Hybrid Matrix Factorization for Recommender Systems in Social
Networks. Neural Network World, 26 (6), 559–569.
〔5〕 Adomavicius, G.; Tuzhilin, A. (2005). Toward the Next Generation of Recommender Systems: A Survey of the State-of-the-Art and Possible Extensions. Knowledge and Data Engineering archive, 17 (6), pp. 734-749
〔6〕 Robin Burke (2007) Hybrid Web Recommender Systems. The Adaptive Web, pp.377-408
〔7〕 Pirasteh P., Jung J.J., Hwang D. (2014) Item-Based Collaborative Filtering with Attribute Correlation: A Case Study on Movie Recommendation. Intelligent Information and Database Systems. ACIIDS 2014, pp. 245-252
〔8〕 V, Subramaniyaswamy; R, Logesh; Chandrashekhar, M; Challa, Anirudh; Varadarajan, Vijayakumar (2017) A personalised movie recommendation system based on collaborative filtering. International Journal of High Performance Computing and Networking 10 (1/2), pp. 54-63
35
〔9〕 Symeonidis, Panagiotis (2007) Content-based Dimensionality Reduction for Recommender Systems. Data Analysis, Machine Learning and Applications March 7-9, pp. 619-626
〔10〕Chumki Basu; Haym Hirsh; William Cohen (1998) Recommendation as Classification: Using Social and Content-Based Information in Recommendations. AAAI '98/IAAI '98 Proceedings of the fifteenth national/tenth conference on Artificial intelligence/Innovative applications of artificial intelligence pp.714-720
〔11〕Suvir Bhargav (2014) Efficient Features for Movie Recommendation Systems. EES Examensarbete / Master Thesis ; XR-EE-KT 2014:012, pp. 53
〔12〕Yiu-Kai Ng (2017) MovReC: A Personalized Movie Recommendation System for Children Based on Online Movie Features. International Journal of Web Information Systems 13 (4): pp. 445-470
〔13〕Abhishek Bhowmick; Udbhav Prasad; Satwik Kottur (2011) Movie Recommendation based on Collaborative Topic Modeling.
〔14〕Shouxian Wei; Xiaolin Zheng; Deren Chen; Chaochao Chen (2016) A hybrid approach for movie recommendation via tags and ratings. Electronic Commerce Research and Applications 18, pp. 83-94

簡易檢索 / 詳目顯示

相關論文