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 1603.09164cs.SI
收藏

蜘蛛和苍蝇:专注于在Tumblr上爬行以检测仇恨促进社区

Spider and the Flies : Focused Crawling on Tumblr to Detect Hate Promoting Communities

Swati Agarwal, Ashish Sureka

摘要:

Tumblr是互联网上最大和最受欢迎的微博客网站之一。研究表明,由于观众的可达性高、出版障碍低和社交网络连通性,微博网站正被现有的极端主义团体滥用为发表仇恨言论和招募新成员的平台。由于每天都有大量的帖子和博客发布,因此手动识别这些帖子和社区是非常不切实际的。我们提出了一个基于主题的网络爬虫主要由多个阶段组成:训练一个文本分类器模型组成的例子,只有仇恨促进用户,提取一个未知的tumblr微博博主的帖子,分类仇恨促进博客的基础上,他们的活动饲料,爬行通过外部链接到其他博客和执行社会网络分析连接极端主义博客。为了研究我们的方法的有效性,我们在大型真实世界数据集上进行了实验。实验结果表明,该方法是一种有效的方法,具有0.80的F-分数。我们应用基于社会网络分析的技术,并确定社区中有影响力的核心博客。

英文摘要:

Tumblr is one of the largest and most popular microblogging website on the Internet. Studies shows that due to high reachability among viewers, low publication barriers and social networking connectivity, microblogging websites are being misused as a platform to post hateful speech and recruiting new members by existing extremist groups. Manual identification of such posts and communities is overwhelmingly impractical due to large amount of posts and blogs being published every day. We propose a topic based web crawler primarily consisting of multiple phases: training a text classifier model consisting examples of only hate promoting users, extracting posts of an unknown tumblr micro-blogger, classifying hate promoting bloggers based on their activity feeds, crawling through the external links to other bloggers and performing a social network analysis on connected extremist bloggers. To investigate the effectiveness of our approach, we conduct experiments on large real world dataset. Experimental results reveals that the proposed approach is an effective method and has an F-score of 0.80. We apply social network analysis based techniques and identify influential and core bloggers in a community.

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。

↑