{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/twitch-plays-pokemon-machine-learns-twitch","title":"Twitch Plays Pokemon, Machine Learns Twitch: Unsupervised Context-Aware Anomaly Detection for Identifying Trolls in Streaming Data","arxiv_id":"1902.06208","date":"2019-02-17","proceeding":null,"authors":["Albert Haque"],"abstract":"With the increasing importance of online communities, discussion forums, and\ncustomer reviews, Internet \"trolls\" have proliferated thereby making it\ndifficult for information seekers to find relevant and correct information. In\nthis paper, we consider the problem of detecting and identifying Internet\ntrolls, almost all of which are human agents. Identifying a human agent among a\nhuman population presents significant challenges compared to detecting\nautomated spam or computerized robots. To learn a troll's behavior, we use\ncontextual anomaly detection to profile each chat user. Using clustering and\ndistance-based methods, we use contextual data such as the group's current\ngoal, the current time, and the username to classify each point as an anomaly.\nA user whose features significantly differ from the norm will be classified as\na troll. We collected 38 million data points from the viral Internet fad,\nTwitch Plays Pokemon. Using clustering and distance-based methods, we develop\nheuristics for identifying trolls. Using MapReduce techniques for preprocessing\nand user profiling, we are able to classify trolls based on 10 features\nextracted from a user's lifetime history.","url_abs":"http://arxiv.org/abs/1902.06208v1","url_pdf":"http://arxiv.org/pdf/1902.06208v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"twitch-plays-pokemon-machine-learns-twitch","repo_url":"https://github.com/ahaque/twitch-troll-detection","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"anomaly-detection","task_name":"Anomaly Detection"},{"task_slug":"clustering","task_name":"Clustering"},{"task_slug":"contextual-anomaly-detection","task_name":"Contextual Anomaly Detection"},{"task_slug":"fad","task_name":"FAD"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}