Implement Code Suite for Fraud Modeling

[advaithsrao]

Go through prior art of ethical techniques around text classification. What can we find and use?
Write optimal code -> preferably a structured data pipeline that does the following,

• Pulls data based on label data logic & runs preprocessing pipeline (Spark?) Label Data for Fraud emails - Enron dataset #1
• Pretrained RoBERTa training code from Pretrained RoBERTa implementation #2
• Testing module from Testing our model and benchmarking performance #3

The above should be done using one python script call for a user.

Note: We will also integrate the ethical codes using ticket #8