Rapid relevance classification of social media posts in disasters and emergencies: A system and evaluation featuring active, incremental and online learning
作者：
Highlights：
• Abstract and precise relevance criteria for emergency services and classifiers.
• Batch learning for relevance classification using precise relevance criteria.
• Active learning for rapid classification during time-critical disasters.
• Incremental learning for real-time classifier quality prediction during labeling.
• Feedback learning allowing users to correct misclassifications reactively.
摘要
•Abstract and precise relevance criteria for emergency services and classifiers.•Batch learning for relevance classification using precise relevance criteria.•Active learning for rapid classification during time-critical disasters.•Incremental learning for real-time classifier quality prediction during labeling.•Feedback learning allowing users to correct misclassifications reactively.
论文关键词：Crisis management,Information overload,Relevance classification,Social media,Supervised machine learning
论文评审过程：Received 8 April 2019, Revised 6 September 2019, Accepted 21 September 2019, Available online 1 October 2019, Version of Record 1 October 2019.
论文官网地址：https://doi.org/10.1016/j.ipm.2019.102132
