A study of multilingual toxic text detection approaches under imbalanced sample distribution

Keyphrases

• XLM-RoBERTa 100%
• Imbalanced Sample Distribution 100%
• Bidirectional Encoder Representations from Transformers 100%
• Toxic Comments 100%
• Text Vectorization 50%
• Translation Language 50%
• Comment Analysis 50%
• Novel Fusion 50%
• Text Digitization 50%
• Pre-training Model 50%

Computer Science

• Language Modeling 100%
• Sample Distribution 100%
• Bidirectional Encoder Representations From Transformers 100%
• Experimental Result 50%
• modeling language 50%
• Preprocessing Step 50%
• Contextual Information 50%
• Training Model 50%
• Translation (Languages) 50%
• Vectorization 50%

Social Sciences

• Multilingualism 100%
• Language Modeling 28%