Detecting Diabetes Using Machine Learning Algorithms

Machine learning has become a significant and promising approach to research and has proven to be an effective technique. It can significantly help in the early diagnosis and prediction of diabetes, which has spread rapidly throughout the world, demanding more modern techniques to detect and diagnose diabetes. In this regard, artificial intelligence (AI) has proven its efficiency in detecting diseases, especially for critical discovery cases depending on a database approved by medical officials. This study estimates the incidence of diabetes using classification techniques that employ a variety of machine learning algorithms, such as Logistic Regression (LR) or so-called Classifier, Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), K-Nearest Neighbor (KNN), and Naive Bays (NB) classifiers. These techniques are used to determine the prevalence of diabetes in all methods related to the preprocessing data, including cleaning data, feature selection, scaling data, and splitting of data. Depending on patient electronic health data (Pima Diabetes Database of India), the present work is implemented using Python programming language within the ANACONDA environment under Jupyter (notebook 6.1.4) and Spyder (4.1.5) applications, and the outcomes show its ability to obtain accurate models with excellent results that can predict the presence of diabetes with up to 100% accuracy.