Convolutional Neural Network (CNN) with word embedding for toxicity analysis in any sentence.
Classify the sentence into different toxicity labels such as ‘toxic’, ‘severe_toxic’, ‘obscene’, ‘threat’, ‘insult’ and ‘identity_hate’ using deep learning model.
Wiki and fast text Dataset used.
Model | Embeddings | AUC | LOSS | Epoch |
---|---|---|---|---|
CNN | Word | 0.98731 | 0.0362 | 2 |