Classifying Arabic Crisis Tweets using Data Selection and Pre-trained Language Models

Arts & Humanities

• Data Selection 100%
• Language Model 80%
• Classifier 34%
• Social Media 32%
• Machine Learning 29%
• Nearest Neighbor 19%
• Tuning 16%
• Situational 15%
• Disaster 13%
• Performance 10%

Social Sciences

• event 14%
• social media 10%
• learning method 6%
• performance 5%
• disaster 5%
• time 2%
• learning 2%