{"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/sequential-prediction-of-social-media","title":"Sequential Prediction of Social Media Popularity with Deep Temporal Context Networks","arxiv_id":"1712.04443","date":"2017-12-12","proceeding":null,"authors":["Bo Wu","Wen-Huang Cheng","Yongdong Zhang","Qiushi Huang","Jintao Li","Tao Mei"],"abstract":"Prediction of popularity has profound impact for social media, since it\noffers opportunities to reveal individual preference and public attention from\nevolutionary social systems. Previous research, although achieves promising\nresults, neglects one distinctive characteristic of social data, i.e.,\nsequentiality. For example, the popularity of online content is generated over\ntime with sequential post streams of social media. To investigate the\nsequential prediction of popularity, we propose a novel prediction framework\ncalled Deep Temporal Context Networks (DTCN) by incorporating both temporal\ncontext and temporal attention into account. Our DTCN contains three main\ncomponents, from embedding, learning to predicting. With a joint embedding\nnetwork, we obtain a unified deep representation of multi-modal user-post data\nin a common embedding space. Then, based on the embedded data sequence over\ntime, temporal context learning attempts to recurrently learn two adaptive\ntemporal contexts for sequential popularity. Finally, a novel temporal\nattention is designed to predict new popularity (the popularity of a new\nuser-post pair) with temporal coherence across multiple time-scales.\nExperiments on our released image dataset with about 600K Flickr photos\ndemonstrate that DTCN outperforms state-of-the-art deep prediction algorithms,\nwith an average of 21.51% relative performance improvement in the popularity\nprediction (Spearman Ranking Correlation).","url_abs":"http://arxiv.org/abs/1712.04443v1","url_pdf":"http://arxiv.org/pdf/1712.04443v1.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":"sequential-prediction-of-social-media","repo_url":"https://github.com/social-media-prediction/flickr-data-prediction-2017","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"prediction","task_name":"Prediction"},{"task_slug":"social-media-popularity-prediction","task_name":"Social Media Popularity Prediction"}],"methods":[],"datasets_introduced":[{"slug":"tpic17","name":"TPIC17","full_name":"Temporal Popularity Image Collection"}],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1712.04443","atlas_url":"https://app.syntology.ai/?focus=1712.04443","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}