Goal:

• Trying to build a text classification model to predict articles based on the hg news dataset
• There are 4 categories [0,3]

Rough idea of steps:

• Import dataset from huggingface, etc.
• Tokenize datasets
  • Tokenize by space
• Create a bag of words for each article
  • Potentially store the bag of words in a tensor, or a hash function + vector, or dictionary
• Perform tf - idf (term frequency - in document frequency)

Bag of words:

• Imagines each article as a “bag of words”
• Stores the frequency of terms from an article in a dictionary
• However, this does not actually