This Jupyter notebook demonstrates how to build an comment classifier using Python and TensorFlow. The classifier distinguishes between 6 types of toxicity level - 'toxic', 'severe_toxic', 'obscene', 'threat', 'insult', 'identity_hate'

1. Downloading data from Kaggle Link - kaggle competitions download -c jigsaw-toxic-comment-classification-challenge.

2. Building a Data Pipeline

   • Constructed a data pipeline using TensorFlow dependencies to manage the flow of image data.

3. Preprocessing text data'comments'

   • Cleaned the text data using a cleaning function- remove puncutaion, remove digits, remve stopwords, made lower case, tokenised, lemmatised.
   • Vectorised the text data.

4. Creating a Deep Neural Network Classifier

   • Used the Sequential library from Keras to build a deep neural network (DNN) classifier.

5. Evaluating Model Performance

   • Assessed the model's performance using standard evaluation metrics - Precision, Recall and Accuracy.

6. Saving the Model for Deployment

   • Saved the trained model to a file for future deployment.

• TensorFlow: For building and training the neural network.
• nltk: For text processing tasks.
• Matplotlib: For plotting and visualisation.
• Keras: Specifically the Sequential API for constructing the neural network.
• OS: For file and directory operations.