Phase 1

Conduct experiments and choose a direction for research. Specifically, Heffan3997 is giving 91% accuracy f1 score on the valid set. Starting phase 2 will continue the directions by finding ways to improve the architecture of Heffan3997:

• For example, replace the 2 - 1 ANN module with other methods such as reducing the dimensionality of some other features
• Organize the code to re-implement the entire Heffan3997 in a more implementable way

Full Changelog: https://github.com/Harito97/Research_ThyroidFNA_ClassAI/commits/pre-release