Classification Report of Distilled Embeddings Classified Using Model Trained with Individual User Post Embeddings. Precision in this context is equivalent to positive predictive value. Recall is equivalent to sensitivity. f1-score is the harmonic mean of both precision and recall. Support is the number of embeddings belonging to a category that were submitted for classification. Categories with f1-scores of less than 0.8 (the categories that are least accurately classified by the model) are denoted with a *.

Similar Items

• Classification Report of Individual User Post Embeddings Classified Using Model Trained with Distilled Post Embeddings. Precision in this context is equivalent to positive predictive value. Recall is equivalent to sensitivity. f1-score is the harmonic mean of both precision and recall. Support is the number of embeddings belonging to a category that were submitted for classification. Categories with f1-scores of greater than 0.6 (the categories that were most accurately classified by the model) are denoted with a *.
  by: Andrew William Alexander (21807525)
  Published: (2025)
• TDA mapping of all distilled embeddings colored by category.
  by: Andrew William Alexander (21807525)
  Published: (2025)
• Confusion Matrix of Distilled Embeddings Classified Using Model Trained with Individual User Post Embeddings.
  by: Andrew William Alexander (21807525)
  Published: (2025)
• Confusion Matrix of Individual User Post Embeddings Classified Using Model Trained with Distilled Post Embeddings.
  by: Andrew William Alexander (21807525)
  Published: (2025)
• Zero-Shot Classification of Hate Speech and Misinformation Distilled Embeddings Using Individual User Post Embeddings as Training Data. Any disorders that were included in the training data but are not listed in this table had neither hate speech embeddings nor misinformation embeddings classified as belonging to them.
  by: Andrew William Alexander (21807525)
  Published: (2025)