Join our book community on Discord
https://packt.link/EarlyAccessCommunity
Most of the data stored in the world is not in the form of numerical data. Just think of the vast information stores in corporate wikis and knowledge bases – not to mention things many of us use every day like email and online message boards. The common connection between all these is that this data is stored as text…and the problem with text-based data is that computers do not inherently understand how to read text. This means that in order to unlock this potential treasure trove of data (which we are already doing a very good job of in the development and utilization of large language models, or large language models (LLMs), like GPT, Llama, and Claude), we need methods to process text into a format digestible by computers. In this chapter, we will cover natural language processing techniques using scikit-learn, including text vectorization, feature extraction, and multiclass classification strategies...
