Codes and Description of our solution for QIA-Hackathon2023 - Prediction of MBTI personality from a set user informations
We also provide the paper for better understanding of novel ideas and techniques used.
our best approaches for each phase, can be found within the title -> embed_phase1_best.ipynb. and train_phase1_best.ipynb. The same is true for phase 2
the other files are trials we have made. The ones which are in the Failures folder are ipynb files which gave results far from our current best, and the ones in the same path as the best files yields closer results to our best performing model.
For implementing traditional ML algorithms we have utilized the cuML library to harness the power of GPU. Here is an instruction how to install it https://docs.rapids.ai/install#get-rapids.