YouTube典型的推薦方法是找尋與使用者點選影片的標題或標籤相似的影片作為推薦；或者先搜尋使用者帳號檔案裡標記喜歡的影片，再推薦與這些喜歡影片相似的影片給使用者。然而此推薦方法有四個缺點，第一是並非每個使用者都有帳號，易造成協同過濾式推薦的冷啟動問題；第二是在當次推薦中易侷限於某標題和標籤而忽略了其他使用者可能感興趣的影片；第三是在連續點選中易重複推薦已推薦過甚至是點選過的影片；第四是如果使用者想鎖定觀看某節目的影片，YouTube易因影片的主題或標籤相似轉而推薦其他節目的影片。
　　本論文提出一個名為MyYouTube的推薦系統，主要希望能為那些只有大方向偏好而沒有明確觀看目標的YouTube使用者，推薦指定節目中多種其可能感興趣且低重複的YouTube影片。在鎖定單一節目的推薦方面，只擷取YouTube中該節目的關聯性影片作為推薦影片資料庫；在提供多種使用者可能感興趣的主題影片方面，採取社交網路中影片和評論者的網路來產生推薦影片清單，並調整連續點選影片之間共同評論者的權重值，使得與使用者選擇越相似的評論者其所看過的影片可以優先被推薦，此外還利用評論者喜好過濾負評過多的影片；在減少影片重複推薦方面，利用使用者的喜好對那些先前推薦過且使用者已瀏覽過其名稱卻沒有點選的影片，給予相當程度的興趣衰減，讓其在後續推薦中不易再出現。
　　實驗結果發現在同一個搜尋關鍵字的影片範圍之內，MyYouTube所產生的前22筆推薦清單比YouTube更能提供多種不同的影片主題。此外，MyYouTube比起YouTube更能有效區分使用者有無瀏覽過的影片名稱，進而減少影片的重複推薦。

The typical recommendation method of YouTube is to search those videos with similar titles or tags that are chosen and marked as favorite videos in a user account. There are four drawbacks in this method. First, not every user has a YouTube account, which may cause a cold-start problem of the collaborative filtering recommender system. Second, this method may easily be restricted to certain titles and tags and therefore neglect other videos in which a user may be interested. Third, this method may easily recommend the same videos due to successively clicking. Fourth, if the user wants to focus on a certain program, it is easy for YouTube to recommend videos of other programs because of the similarity of the titles or tags.
This thesis presents a recommendation system called MyYouTube. It is mainly to recommend various and lowly repeated videos for those users who have preference about certain program but no specific watching objective. In the aspect of recommending a single program, the system simply extracts the related videos of the program from YouTube as it is very straightforward to search a specific program. In the aspect of providing various topic videos, the system adopts the network between the videos and the reviewers from YouTube to generate a recommendation watching list and also reweigh the co-reviewers between the videos which are successively being clicked. By doing this way, the videos can be recommended in priority if their reviewers have similar choices as the user. The system can also filter the videos that have many negative marks by using the preference of the reviewers. In the aspect of reducing repeated recommendation, the system uses a user's preference to decrease the weights of videos which have been previously recommended and browsed, but not elected by the user.
Our experimental result demonstrates that, with same searching keywords, the proposed MyYouTube recommends more videos with various topics than YouTube in the first 22 recommendation items. In addition, MyYouTube can distinguish more efficiently than YouTube the titles of the videos of having being browsed or not by the user to reduce the repetition of recommending videos.

[1] YouTube. http://www.youtube.com .
[2] comScore, Inc. http://www.comscore.com .
[3] P. Resnick, N. Iacovou, M. Suchak, P. Bergstorm, and J. Riedl, "GroupLens: An open architecture for collaborative filtering of netnews," in Proceedings of the 1994 ACM international conference on Computer Supported Cooperative Work, North Carolina, USA, pp. 175-186, 1994.
[4] D. Goldberg, D. Nichols, B. M. Oki, and D. Terry, "Using collaborative filtering to weave an information tapestry," in Proceedings of ACM international journal on Communications, vol. 35, no. 12, pp. 61–70, 1992.
[5] U. Shardanand and P. Maes, "Social information filtering: Algorithms for automating ‘word of mouth’," in Proceedings of ACM international conference on Human Factors in Computing Systems, Denver, Colorado, USA, pp. 210-217, May 1995.
[6] W. Hill, L. Stead, M. Rosentein, and G. W. Furnas, "Recommending and evaluating choices in a virtual community of use," in Proceedings of ACM conference on Human Factors in Computing Systems, New York, USA, pp. 194–201, 1995.
[7] G. Linden, B. Smith, and J. York, "Amazon.com recommendations: item-to-item collaborative filtering," in Proceedings of IEEE international journal on Internet Computing, vol. 7, no. 1, pp. 76-80, February 2003.
[8] J. A. Alspector, A. Kolcz, and N. Karunanithi, "Comparing feature-based and clique-based user models for movie selection," in Proceedings of the Third ACM international conference on Digital Libraries, Pittsburgh, Pennsylvania, USA, pp. 11-18, June 1998.
[9] X. S. Hua, M. Worring, and T. S. Chua, Internet Multimedia Search and Mining, Bentham Science Publishers, pp. 1-17, 2010.
[10] G. Adomavicius and A. Tuzhilin, "Toward the Next Generation of Recommender Systems: A Survey of the State-of-the-Art and Possible Extensions," in Proceedings of IEEE journal on Knowledge and Data Engineering, vol.17, no.6, pp. 734-749, June 2005.
[11] Q. Li and B. M. Kim, "Clustering approach for hybrid recommender system," in Proceedings of the IEEE/WIC international conference on Web Intelligence, South Korea, pp. 33-38, October 2003.
[12] M. Balabanovic and Y. Shoham, "Fab: content-based, collaborative recommendation," in Proceedings of ACM international journal on Communications, vol. 40, no. 3, pp. 66-72, March 1997.
[13] A. M. A. Wasfi, "Collecting user access patterns for building user profiles and collaborative filtering," in Proceedings of the 4th ACM international conference on Intelligent User Interfaces, New York, NY, USA, pp. 57-64, 1999.
[14] M. V. Setten and M. Veenstra, "Prediction strategies in a TV recommender system- method and experiments," in Proceedings of IADIS international conference on WWW and Internet 2003, Algarve, Portugal, pp. 203-210, 2003.
[15] J. Bollen, M. L. Nelson, R. Araujo, and G. Geisler, "Video recommendations for the open video project," in Proceedings of the ACM/IEEE-CS joint conference on Digital libraries, Denver, Colorado, USA, pp. 369–369, 2005.
[16] C. C. Tsai, C. I. Chung, Y. T. Huang, C. H. Shen, Y. C. Wu, and J. C. Yang, "VCSR: Video content summarization for recommendation," in IEEE international conference on Advanced Learning Technologies, pp. 862–864, July 2007.
[17] H. Luo, J. Fan, and D. A. Keim, "Personalized news video recommendation," in Proceeding of ACM international conference on Multimedia, Vancouver, British Columbia, Canada, pp. 1001–1002, 2008.
[18] T. Mei, B. Yang, X. Hua, L. Yang, S. Yang, and S. Li, "VideoReach: an online video recommendation system," in Proceedings of the 30th annual international ACM SIGIR conference on Research and Development in Information Retrieval, New York, NY, USA, pp. 767–768, 2007.
[19] A. Krishnakumar, "Recoo: A Recommendation System for Youtube RSS Feeds," Technical Report, UC Santa Cruz IRKM, CA, USA, 2007.
[20] S. Baluja , R. Seth , D. Sivakumar , Y. Jing , J. Yagnik , S. Kumar , D. Ravichandran , M. Aly, "Video suggestion and discovery for YouTube: taking random walks through the view graph," in Proceeding of the 17th ACM international conference on World Wide Web, Beijing, China, pp. 895-904, April 2008.
[21] S. Qin, R. Menezes, and M. Silaghi, "A Recommender System for YouTube Based on its Network of Reviewers," in Proceedings of the 2010 IEEE second international conference on Social Computing, Minneapolis, Minnesota, USA, pp.323-328, 2010.
[22] Google Inc., "Youtube data API," http://code.google.com/intl/en/apis/youtube/overview.html .
[23] J. Davidson, B. Liebald, J. Liu, P. Nandy, T. V. Vleet, U. Gargi, S. Gupta, Y. He, M. Lambert, B. Livingston, and D. Sampath, "The YouTube video recommendation system," in Proceedings of the fourth ACM conference on Recommender Systems, New York, NY, USA, pp.293-296, 2010.
[24] M. Cha, H. Kwak, P. Rodriguez, Y.-Y. Ahn, and S. Moon, "I Tube, You Tube, Everybody Tubes: Analyzing the World’s Largest User Generated Content Video System," in Proceedings of the 7th ACM SIGCOMM conference on Internet Measurement, pp. 1-14, San Diego, CA, USA, October 2007.
[25] G. A. Miller, "The magical number seven, plus or minus two: Some limits on our capacity for processing information," Psychological Review, vol. 63, no. 2, pp.81-97, 1956.