An Improved Hybrid Approach for Coronary Heart Disease Diagnosis Using PSO and GA

The most deadly disease in the world is coronary heart disease. Early detection of cardiac disease is essential as it is the only way to reduce the overall death rate caused by the disease. The majority of methods for diagnosing heart disease are clinical and expert dependent, which renders them expensive and out of reach for the general public. In this era of digital health, numerous researchers and medical professionals have created a variety of computer-aided diagnosis systems by using various machine learning techniques: K-nearest neighbour, random forest, linear regression, naive bayes, neural network, support vector machine, and decision tree are a few examples. This paper proposes a highly automated, less expensive and timesaving diagnosis model of coronary heart disease using fuzzy logic system optimized by particle swarm optimization and genetic algorithm. The model has been applied to the heart patients dataset extracted from Kaggle with thirteen clinical factors. In the initial stage, the system makes use of fuzzy logic with mamdani type-1. Further, the study tunes and optimizes the rules and membership functions using two evolutionary algorithms and compare the results of these techniques through the experimental records, RMSE and cross-validation method. The experimental findings yield 0.3437 and 0.3547 RMSE along with validation of testing set having RMSE 0.367 and 0.369 using particle swarm optimization and genetic algorithm respectively.