arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1412.1185cs.SIcs.CYphysics.soc-ph
收藏

YouTube视频中的注意熵和流行度

The Entropy of Attention and Popularity in YouTube Videos

Jonathan Scott Morgan, Iman Barjasteh, Cliff Lampe, Hayder Radha

摘要:

绝大多数YouTube视频从未流行起来,默默无闻,观看次数很少,没有喜欢,也没有评论。我们使用基于熵的信息理论测量来研究YouTube的“趋势视频”和“最新”视频源中常见的流行度测量的时间序列分布如何与注意力的理论概念相关。虽然“最新”提要中的大多数视频从来都不受欢迎,但其中约20%的视频具有与“趋势视频”的分布类似的关注度量和熵度量的分布。我们分析了20%的“最近”视频变得有点流行与80%的不流行的不同,然后将这些流行的“最近”视频与“趋势视频”的不同子集进行比较,以尝试描述和比较每个视频的关注度。

英文摘要:

The vast majority of YouTube videos never become popular, languishing in obscurity with few views, no likes, and no comments. We use information theoretical measures based on entropy to examine how time series distributions of common measures of popularity in videos from YouTube's "Trending videos" and "Most recent" video feeds relate to the theoretical concept of attention. While most of the videos in the "Most recent" feed are never popular, some 20% of them have distributions of attention metrics and measures of entropy that are similar to distributions for "Trending videos". We analyze how the 20% of "Most recent" videos that become somewhat popular differ from the 80% that do not, then compare these popular "Most recent" videos to different subsets of "Trending videos" to try to characterize and compare the attention each receives.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑