Abstract
Heart disease is emerging as the single most critical reason of mortality worldwide and is one of the costliest chronic conditions. Regardless of heart diseases damaging complications, it is the most preventable and controllable disease; therefore, it is important to predict it ahead of time. Considering the mortality rate of heart disorder and its rising health care costs, it is important to predict this malignant disease at its earliest. There are existing cardiac disorder risk assessment models however they are costly and operationally complex that restrains their use in rural areas and at public-level screening evaluations. To overcome these drawbacks of the prevailing heart disease risk systems, we develop a heart disease risk assessment model that can be utilized for public-standard screening to recognize patients at a high risk of heart disease and produce knowledge to facilitate initial intervention and enhance patient's health. The developed risk assessment model uses random forest, support vector machine and decision tree machine learning techniques. The developed risk model's efficiency is checked using various model and medical metrics. Experimental results show that the random forest risk assessment model outperforms other proposed risk models with the highest recognition rate, precision, sensitivity and AUROC score of 90.42%, 91.97%, 89.75% and 94%. As per our knowledge the experimental results obtained are highest than the published values in the literature. The developed risk model is applicable in where people lack the facilities of the integrated primary medical care technologies.