Navigating Pathways to Automated Personality Prediction

This repository accompanies the paper: Navigating Pathways to Automated Personality Prediction: A Comparative Study of Small and Medium Language Models (Frontiers in Big Data, 2024).

🚀 Overview

The project explores the efficiency of small and medium-sized NLP models in predicting Big Five personality traits from textual data. It emphasizes computational trade-offs, accuracy, and sustainable AI practices in marketing and consumer analytics.

🔍 Methodology

Data Preprocessing: Cleaning and tokenizing text samples for input.

Models Used:

RoBERTa (medium) ALBERT (small, efficient)

Training: Fine-tuned Hugging Face Transformer models with PyTorch. Evaluation Metrics: RMSE, MSE, Loss, computational cost (time & memory).

Comparative Analysis: Benchmarked performance vs. resource consumption to assess sustainable AI choices.

📊 Results

Medium models (e.g., RoBERTa) achieved slightly higher accuracy but required significantly more resources. Smaller models (e.g., ALBERT, DistilBERT) offered competitive accuracy with lower computational costs, making them viable for scalable business applications.

📦 Tech Stack

Python Hugging Face Transformers PyTorch Scikit-learn Pandas & NumPy

🌍 Applications

Marketing & Consumer Research Automated Personality Prediction Scalable AI Deployment

📜 Citation

If you use this work, please cite: Habib, F. (2024). Navigating Pathways to Automated Personality Prediction: A Comparative Study of Small and Medium Language Models. Frontiers in Big Data.