In this paper, we propose a unique hybrid model that combines the Bayesian and Random Forest algorithms to better predict diabetic heart disease. The hybrid model also includes hyperparameter adjustment, model evaluation, and feature selection using SVM weights. The suggested model beats individual models in terms of accuracy and efficiency, according to experimental findings. In contrast to the Random Forest technique, which generates an ensemble of decision trees to increase prediction accuracy, the Bayesian approach offers interpretability and the ability to explain. The suggested model outperformed the individual models, achieving an accuracy of 95% with a precision of 0.90 and recall of 0.88. Python can be used to implement the hybrid model, and we think this model has a lot of potential for enhancing diagnostics.

Bayesian, Random Forest algorithm, diabetic heart disease, prediction, hybrid model.