A comparative analysis of early stage Diabetes Mellitus prediction using machine learning approach

Diabetes mellitus (DM) is a serious chronic illness and causes a severe impact on the complete human body. Diabetes that goes undiagnosed increases the risk of heart attack, renal disease, eye problems, and organ failure. Many diabetic patients are unaware that they have diabetes, and its complications are rising day by day. As Diabetes mellitus is not curable, it must be detected early to live a healthy life. The purpose of this research is to use relevant features to create a machine learning model and choose the best classifier to get the best results when compared to clinical results. The Diabetes dataset is taken from Kaggle. The dataset is preprocessed to improve the model’s learning capabilities. To identify individuals as diabetes or non-diabetic, several machine learning approaches such as Logistic Regression, KNeighborsClassifier, Naive Bayes, Random Forest classifier, Decision tree classifier, AdaBoost, and XGBoost, Light GBM is used. To choose the best model, the performance of machine learning models is compared by using accuracy, recall, precision, and F1-Score as the evaluation metrics.Finally got an accuracy of 99% in DecisionTree.