The Characteristics of Youku Videos and User Intentions

• Main Authors: Kai-Hung Cheng, 鄭凱鴻
• Other Authors: Bang Ye Wu
• Format: Others
• Language: en_US
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/j59m87

碩士 === 國立中正大學 === 資訊工程研究所 === 102 === With thousands of videos submitted every day, Youku has become the biggest user generated content (UGC) site in China. Once registered, users are able to upload videos irrespective of video length, comment and rate other videos. In this thesis, our objective is to find correlations among several attributes and analyze the major factor that impact view counts of videos. Besides, we also discovered several characteristics of Youku videos through our sampling datasets. In our thesis, we conjecture that most of the users tend to treat Youku as a video source, rather than a social network site. The above factors inspire us to analyze the major factor that impacts view count most subsequently. To this end, we analyze Pearson correlations between view count and other attributes and find that it is the fans number a user have that impacts his/her view count most. The result reflects users’ behavior. Users tend to follow someone whose content of videos fit their interest. We also analyze two rankings Youku provides every day and find that most top ranking videos will last no more than 3 days, yet videos of other categories such as movies may last longer. To discover whether the search frequency of a term affect its view counts, we also evaluate the rank correlation between the two rankings, and conclude that search counts usually have less to do with view counts.