MGEL: Multigrained Representation Analysis and Ensemble Learning for Text Moderation

Keyphrases

• Abusive Language 33%
• Art Performance 33%
• ASCII Character 33%
• Bidirectional Encoder Representations from Transformers 33%
• Byte 66%
• Character Decomposition 33%
• Character-level 33%
• Classification Methods 33%
• Comprehensive Experiment 33%
• Context Understanding 33%
• Ensemble Learning 100%
• Level Scheme 33%
• Logistic Regression 33%
• N-gram 33%
• Neural Network 33%
• Obfuscation 100%
• One-hot 33%
• Over-representation Analysis 100%
• Popular Texts 33%
• Text Classification 33%
• Weighting Approach 33%

Computer Science

• Art Performance 33%
• Bidirectional Encoder Representations From Transformers 33%
• Classification Method 33%
• Ensemble Learning 100%
• Logistic Regression 33%
• Multiplicity 33%
• Neural Network 33%
• Obfuscation 100%
• Text Classification 33%