Tuning Hyper Parameters for Predicting Coronary Heart Disease Based on Grid Search Algorithm

Heart disease proves to be the primary reason behind both death and disability. Prognosis in the early stages with adequate accuracy is essential for preventing and managing coronary artery disease (CAD) effectively. But that's not enough. The idea of this paper is a cardiac disease prediction ML model that involves numerous preprocessing steps and optimization techniques in terms of hyperparameters. This type of heart illness can be anticipated to improve the precision in the diagnosis of cardiac problems. Therefore, applying Grid Search Optimization (GSO) ML models is a feasible way to predict the disease promptly. In order to reduce the false-negative rate, the most advanced approach simultaneously incorporates feature selection and hyperparameter adjustment in the search for the model. Three models complete the task with cross-validation. Analysis of statistical and correlation matrices for feature selection in multivariate regression. It was possible to get the outcomes of the machine learning methods' performances. At 90.14 percent, the Random Forest (RF) model with fine-tuned hyperparameters was the most accurate in illness prediction. This study made use of the Z-Alizadeh sani dataset. The best parameters for prediction were found by analyzing the experimental results. This work's technique significantly outperformed previous research.