Predicting Diabetes u sing SVM Implemented by Machine Learning

Abstract

Age, BMI, and insulin levels, which play important roles because they are not constant and do not follow any specific patterns, are some of the factors that can be used to identify the chronic disease of Diabetes. Besides the elements described above, a few additional will be studied in subsequent subjects in this study. Before cleaning the data, support vector machine (SVM) algorithms, pandas, NumPy, and sci-kit-learn libraries are used to predict the patient's diagnosis and classify the data into various categories. The output contains two parameters: DIABETIC and NON-DIABETIC. With the available dataset, the accuracy score of training data was 77.5 percent and the accuracy score of test data was 80.5 percent.