Comparación de modelos de NLP para la extracción de entidades nombradas en acciones de tutelas asociadas a temas de salud

Abstract

The Colombian healthcare system is often affected by issues such as corruption, inefficiency, and delays in administrative processes, which frequently result in user complaints. These complaints often infringe on individuals' fundamental rights, leading to the filing of legal actions to claim such rights. This document proposes extracting relevant entities for the classification and categorization of legal actions using various natural language processing models. The performance of these models will be evaluated using specialized metrics for named entity extraction, allowing for a comparison of their effectiveness. Additionally, adjustments to hyperparameters and training methods will be made to determine the most suitable model for this task.