KM0926/bigfivepredictions_pandora

This repository includes the models for a research project focused on Personality Trait Prediction using four models: BERT base, BERT large, RoBERTa base and RoBERTa large. The objective is to investigate how well these models predict personality traits and identify correlations among them based on user-generated text collected from Reddit (arXiv:2004.04460v3). Unlike the models used in the multiple-models approach, the models for the single-model approach are not included separately, as the final RoBERTa base model was directly used for evaluation and comparison in the single-model setting.
The full paper of this project can be found at the following link: https://www.mdpi.com/2078-2489/16/5/418

The models are organised into the following modules:
bertbase_best_model : Best BERT base model saved during the fine-tuning process
bertbase_final_model : Final BERT base model saved after the fine-tuning process
bertlarge_best_model : Best BERT large model saved during the fine-tuning process
bertlarge_final_model : Final BERT large model saved after the fine-tuning process
robertabase_final_model : Final RoBERTa base model saved after the fine-tuning process
robertalarge_best_model : Best RoBERTa large model saved during the fine-tuning process
robertalarge_final_model : Final RoBERTa large model saved after the fine-tuning process
r_a_best_model : Best RoBERTa base model for Agreeableness saved during the fine-tuning process for the multiple-models approach
r_a_final_model : Best RoBERTa base model for Agreeableness saved during the fine-tuning process for the multiple-models approach
r_c_best_model : Best RoBERTa base model for Conscientiousness saved during the fine-tuning process for the multiple-models approach
r_c_final_model : Best RoBERTa base model for Conscientiousness saved during the fine-tuning process for the multiple-models approach
r_e_best_model : Best RoBERTa base model for Extraversion saved during the fine-tuning process for the multiple-models approach
r_e_final_model : Best RoBERTa base model for Extraversion saved during the fine-tuning process for the multiple-models approach
r_n_best_model : Best RoBERTa base model for Neuroticism saved during the fine-tuning process for the multiple-models approach
r_n_final_model : Best RoBERTa base model for Neuroticism saved during the fine-tuning process for the multiple-models approach
r_o_best_model : Best RoBERTa base model for Openness saved during the fine-tuning process for the multiple-models approach
r_o_final_model : Best RoBERTa base model for Openness saved during the fine-tuning process for the multiple-models approach

*Note: The final model for RoBERTa base was also the best model; therefore, no separate best checkpoint was saved.