Implementation of a Prediction System for Heart Failure Mortality Using an Artificial Intelligence Model

Espinoza R.; Salvador C.; Añez J.; Huaman A.; Ticona W. · Proceedings of the 14th International Conference on Cloud Computing, Data Science and Engineering, Confluence 2024 (2024) · DOI: 10.1109/confluence60223.2024.10463319

Resumen

Today heart failure diseases affect people around the world, the WHO estimates that this disease affects annually around 26 million people worldwide and in turn is responsible for numerous hospitalizations and deaths. Therefore, the study aims to find the best artificial intelligence model for the prediction of mortality in patients suffering from heart failure. For this, a database of 299 patients with heart failure disease has been used, where characteristics of each of them have been collected based on their clinical history. The MinMax data normalization technique has been used to standardize the values of the characteristics. Finally, different Machine Learning models were developed for the prediction of patient mortality, which were, Support Vector Machine, Decision Tree, and Random Forest. The results were evaluated based on the metrics of Accuracy, accuracy, Recall and F1-score. The model that obtained the best results was Random Forest obtaining 92% accuracy, Recall 86% and F1-score 84%.

Instituciones: Universidad ESAN; Universidad Tecnológica del Perú