作者: Michalis Faloutsos , Evangelos E. Papalexakis , Joobin Gharibshah , Jakapun Tachaiya
DOI: 10.1109/ASONAM49781.2020.9381312
关键词:
摘要: Online forums have been shown to contain a wealth of useful information. With few notable exceptions, such not received much attention from the research community, unlike other online social media. Our goal here is conduct an in-depth thread-centric analysis forums, focusing on security forums. We propose, RThread, comprehensive unsupervised clustering approach with powerful visualization component, which we provide as publicly-accessible web-based tool. leverages 92 thread features that span three groups: (a) temporal, (b) behavioral, and (c) content related. analyze data 8 400k posts over years. First, find many properties follow log-normal distribution, persistent across several time. Second, show how our can identify clusters threads similar behavior, while component provides easy way spot differences between these clusters. Finally, surprising behaviors, including cluster, whose are used for Search Engine Optimization. see publicly available platform building block towards understanding forum activity extracting interesting information in way.