A neural network designed to take in plain text and classify it’s emotions. It classifies text into one of 6 categories:

  • Toxic
  • Severely Toxic
  • Obscene
  • Threat
  • Insult
  • Identity Hate

The data was gathered from a Kaggle competition, linked Here.

What I Learned

  • Perform data cleanup and analysis with tools such as NumPy and Pandas
  • Create deep learning neural networks using popular Python libraries such as Sckit-Learn
  • Perform multiple epochs, adjusting the activation layers and network properties to achieve the most accurate result
  • Utilize vectorization to process plain text, requiring extensive linear algebra calculations