/Automatic-Cyberbullying-Detection

PolEval 2019 Task6-1. Harmful vs Non-Harmful.

Primary LanguageJupyter Notebook

Automatic-Cyberbullying-Detection

PolEval 2019 Task6-1 (link). Harmful vs Non-Harmful (binary classification).

GoogleColab

Click: Open In Collab

Used models:

  • SpaCy TextCategorizer
  • Linear Regression (optimized with GridSearchCv)

Used performance measures:

  • Accuracy
  • Precision
  • Recall
  • F1-Score

PolEval data distribution:

poleval_data_distribution.png