XAI-Augmented Voting Ensemble Models for Heart Disease Prediction: A SHAP and LIME-Based Approach

Bioengineering, Journal Year: 2024, Volume and Issue: 11(10), P. 1016 - 1016

Published: Oct. 12, 2024

Ensemble Learning (EL) has been used for almost ten years to classify heart diseases, but it is still difficult grasp how the “black boxes”, or non-interpretable models, behave inside. Predicting disease crucial healthcare, since allows prompt diagnosis and treatment of patient’s true state. Nonetheless, forecast illness with any degree accuracy. In this study, we have suggested a framework prediction based on Explainable artificial intelligence (XAI)-based hybrid such as LightBoost XGBoost algorithms. The main goals are build predictive models apply SHAP (SHapley Additive expPlanations) LIME (Local Interpretable Model-agnostic Explanations) analysis improve interpretability models. We carefully construct our systems test different ensemble learning algorithms determine which model best (HDP). approach promotes transparency when examining these widespread health issues. By combining XAI, important factors risk signals that underpin co-occurrence made visible. accuracy, precision, recall were evaluate their efficacy. This study highlights healthcare be transparent recommends inclusion XAI medical decisionmaking.

Language: Английский

Predictive value of heart rate for prognosis in patients with cerebral infarction without atrial fibrillation comorbidity analyzed according to the MIMIC-IV database

Frontiers in Neurology, Journal Year: 2025, Volume and Issue: 16

Published: March 14, 2025

This study focused on the relationship between heart rate and likelihood of death within 28 days in patients with cerebral infarction without comorbidity atrial fibrillation, using patient data extracted from MIMIC-IV database. involved a retrospective analysis clinical 1,643 individuals who were admitted to ICU. To investigate role determining survival, we applied variety statistical techniques such as Cox regression models, survival Kaplan-Meier plots, spline-based models. In addition, performed analyses by subgroups identify any potential variables that could influence association HR 28-day mortality. univariate multivariate analyses, elevated was strongly associated higher mortality, even after adjusting for confounders age, sex, comorbidities, scores.(HR:1.01, 95%,CI:1.01 ~ 1.02, p = 0.019) showed > 90 beats/min had significantly lower probability survival. Restricted cubic spline (RCS) confirmed nonlinear Subgroup demonstrated an interaction factors hypertension mechanical ventilation status. highlights prognostic significance independent predictor mortality do not have fibrillation.

Language: Английский

Utilizing Voting Classifiers for Enhanced Analysis and Diagnosis of Cardiac Conditions

Results in Engineering, Journal Year: 2025, Volume and Issue: unknown, P. 104636 - 104636

Published: March 1, 2025

Language: Английский