Detecting hashtag hijacking from Twitter
Courtland VanDam, Pang‐Ning Tan · 2016
Twitter hashtags are typically used to categorize a tweet, to monitor ongoing conversations, and to facilitate accurate retrieval of posts. Hashtag hijacking occurs when a group of users starts using a trending hashtag to promote a topic that is substantially different from its recent context. Most of the prior research on hashtag hijacking has focused on manual monitoring of specific hashtags. We present a general framework based on multi-modal matrix factorization for automatically detecting hashtag hijacking from Twitter data, where the compromised hashtags and their underlying topics were unknown a priori.