{"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/learning-reporting-dynamics-during-breaking","title":"Learning Reporting Dynamics during Breaking News for Rumour Detection in Social Media","arxiv_id":"1610.07363","date":"2016-10-24","proceeding":null,"authors":["Arkaitz Zubiaga","Maria Liakata","Rob Procter"],"abstract":"Breaking news leads to situations of fast-paced reporting in social media,\nproducing all kinds of updates related to news stories, albeit with the caveat\nthat some of those early updates tend to be rumours, i.e., information with an\nunverified status at the time of posting. Flagging information that is\nunverified can be helpful to avoid the spread of information that may turn out\nto be false. Detection of rumours can also feed a rumour tracking system that\nultimately determines their veracity. In this paper we introduce a novel\napproach to rumour detection that learns from the sequential dynamics of\nreporting during breaking news in social media to detect rumours in new\nstories. Using Twitter datasets collected during five breaking news stories, we\nexperiment with Conditional Random Fields as a sequential classifier that\nleverages context learnt during an event for rumour detection, which we compare\nwith the state-of-the-art rumour detection system as well as other baselines.\nIn contrast to existing work, our classifier does not need to observe tweets\nquerying a piece of information to deem it a rumour, but instead we detect\nrumours from the tweet alone by exploiting context learnt during the event. Our\nclassifier achieves competitive performance, beating the state-of-the-art\nclassifier that relies on querying tweets with improved precision and recall,\nas well as outperforming our best baseline with nearly 40% improvement in terms\nof F1 score. The scale and diversity of our experiments reinforces the\ngeneralisability of our classifier.","url_abs":"http://arxiv.org/abs/1610.07363v1","url_pdf":"http://arxiv.org/pdf/1610.07363v1.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":"learning-reporting-dynamics-during-breaking","repo_url":"https://github.com/yunzhusong/aard","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"learning-reporting-dynamics-during-breaking","repo_url":"https://github.com/znhy1024/heard","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"rumour-detection","task_name":"Rumour Detection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1610.07363","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}