Computer vision-based predictive analysis of chronic cardiovascular disease using heartbeat features

Heart disease has become one of the world’s most dangerous and serious diseases due to the difficulty in identifying it. Machine learning may assist the medical field by providing accurate and timely illness diagnoses. The main objective is to develop an expert system that can identify heart illness early and assist cardiologists in making more accurate diagnoses. This paper examines many aspects of heart illness and develops a model using supervised learning techniques like Logistic Regression, Naïve Bayes, decision trees, K-nearest neighbor (KNN), random forest, Support Vector Machine (SVM), and the XG Boost methods. It utilizes a dataset of the Cleveland catalog of heart patients at UCI. There are around 303 occurrences plus 76 characteristics within the collection. Only 14 out 76 characteristics are tested, despite their importance in proving the effectiveness of alternative algorithms. The accuracy of different classifiers are,in Naive Bayes 83.51%, SVM 84.79%, Decision Tree 77.5%, Random Forest 87.94%, and KNN 80.21%, Logical Regression 85.0% is obtained. Gradient Boost has the highest accuracy rate of 95.62%, Sensitivity of 92.3%, Precision 83.7%, and specificity of 93.8 Percent in the case of same data set we have applied for other classifier.