Final

• Firstly, big thanks to the users who provided the notebooks above. This notebook is merely adding some utility code to make a solid baseline that other users may iterate on, but all the heavy lifting was done by the above notebooks.
• I encapsulated the analyzer in a class, and added code to run the analyzer on (potentially) both the training and test set. I also added validation code, so that we can analyze the performance of the analyzer on the training set.
• I also added a global configuration for ease of testing, which allows the user to switch between training and inference mode. Additionally, I incorporated the external data that https://www.kaggle.com/alejopaullier kindly provided in his discussion post.
• For now, the business logic roughly stays the same as the Modify of PII Detect Study notebook that I used beforehand.