Predicting Diabetes in Bangladesh Using Machine Learning: A Data-Driven Approach

Diabetes is a chronic condition caused by inadequate insulin synthesis, which prevents the body fromprocessing blood sugar. Diabetes' etiology is still mostly unknown. Furthermore, diabetes is not routinelymeasured. Since there is no known treatment for diabetes, it is crucial for those who have disease to monitortheir blood sugar levels in order to manage and maintain their health. Therefore, safeguards against diabetesor early detection are necessary. Diabetes symptoms can occur suddenly, and it can be mild so it can take yearsto notice. Diabetes can damage the heart, eyes, kidneys, nerves and damage blood vessels over time. As of today,no known means to prevent diabetes nor its cause are known so early diagnosis or predicting diabetes isnecessary to prevent the worst effects of diabetes. Additionally, People in Bangladesh are still ignorant aboutdiabetes and unaware exactly when diabetes has to be measured. In this paper, we are going to predict diabetesby using machine learning. We compared conventional machine learning with deep learning approaches. Forthe conventional machine learning method, we considered the most commonly used classifiers: K-NearestNeighbors (KNN) and Random Forest (RF). On the other hand, for Deep Learning (DL) we employed a fullyArtificial Neural Networks (ANN) and Long Short-Term Memory (LSTM) Algorithm to predict and detectdiabetes patients.