Prediction of Heart Attack Risk with Data Mining by Using Blood Tests and Physical Data

The aim of this study was to more accurately predict the presence of heart disease using 12 characteristics: cholesterol, glucose, systolic blood pressure, diastolic blood pressure, smoking, alcohol, age, height, weight, gender, physical activity, and cardiovascular disease status. Naive Bayes, J48 Decision Tree, and Function Simple Logistic Algorithms were applied to the dataset with the WEKA program. The algorithms were applied to the Kaggle dataset (70,000 records, 11 features, and a target), and the accuracy rates of these data mining classification algorithms were compared. Data cleaning (getting rid of, and getting rid of outliers with IQR methods), data processing and forecasting, and supervised machine learning (using a training set and test set) were used in this study. According to the results, it is seen that the Naive Bayes algorithm is the most accurate and gives the best results. The J48 algorithm’s result is %87, and the Function Simple Logistic Algorithm’s result is %86. Our Naive Bayes algorithm’s score is %89, which is higher than the literature studies.