A hybrid approach with metaheuristic optimization and random forest in improving heart disease prediction.

Cardiovascular diseases (CVD)  a major cause of morbidity and mortality among the world's non-communicable disease incidences. Though these practices are in use for diagnostics of different CVDs in clinical settings, need improvement because they are solving the purpose of only 57% of the patients in emergency. Due to this cost of diagnosis for heart disease is increasing which is the reason for analyzing heart disease and predicting it as early as possible. The main motive of this paper is to find an intelligent method for predicting disease effectively by means of machine learning (ML) and metaheuristic algorithms. Optimization techniques have the merit of handling non-linear complex problems. In this paper, an efficient ML model along with metaheuristic optimization techniques is evaluated for heart disease dataset to enhance the accuracy in predicting the disease. This will help to reduce the death rate due to the severity of heart disease. The SelectKBest feature selection is applied to the Cleveland Heart dataset and overall rank is obtained. Accuracy is measured. The optimization techniques namely Genetic Algorithm Optimized Random Forest (GAORF), Particle Swarm Optimized Random Forest (PSORF), and Ant Colony Optimized Random Forest (ACORF) are applied to the Cleveland dataset. Classification algorithms are performed before and after optimization. The output of the experiment explains that the GAORF performed better for the dataset considered. Also, a comparison is made along with the SelectKBest filter methods. The proposed model achieved better accuracy which is the maximum among other optimization and classification techniques.