Data Extraction, Visualization, and Prediction Through Natural Language Processing

Alvarado C.; Velásquez G.; Mauricio D. · 2024 IEEE International Conference on Omni-Layer Intelligent Systems, COINS 2024 (2024) · DOI: 10.1109/coins61597.2024.10622130

Resumen

This study presents Datalyzer, a system designed for data extraction, visualization, and prediction in the mining sector using advanced NLP and machine learning, specifically GPT-3.S Turbo. The system enhances operational efficiency through rigorous data preprocessing and specialized fine-tuning, validated on a simulated mining dataset. Results show significant improvements: data extraction time reduced by 94 % and visualization time by 97.6%. These improvements indicate a transformation in efficiency, usability, and user satisfaction. Despite limitations in data variability and complexity, this pioneering approach highlights the potential of NLP and machine learning in modernizing the mining industry and supporting data-driven decision-making.

Instituciones: Universidad Peruana de Ciencias Aplicadas