A Predictive Diagnostic Model for Diabetes Using Machine Learning Technique

Diabetes is a primary chronic disease caused by a succession of metabolic abnormalities in which blood glucose levels are abnormally high for an unspecified amount of time. It affects every organ in the human body, leading to a wide range of complicated illnesses like vision, renal disease, pulmonary embolism, stroke, and so forth. Diabetes ailments are currently among the leading causes of death in healthcare. The possibility and severity of diabetes may be considerably reduced if a reliable early diagnosis is made. Machine learning (ML) techniques are utilized to analyze medical information at an earlier stage of life to keep individuals safe. In this research, early diabetes prediction and determining the best-performing classifier is a significant concern. To predict diabetes in a patient, different machine learning classification algorithms, such as random forest, K-nearest neighbour (KNN), decision tree (DT), and logistic regression were employed on the PIMA Indian Diabetes Dataset (PIDD) to predict diabetes in patients. The evaluation of the performance of all the classification methods employed was done with various measurement methods such as accuracy, precision, and F1-score. The performance of the various ML algorithms employed in this research suggests the algorithm is most suitable for diabetes prediction. It is observed that among all the models logistic regression outperformed the other ML techniques with an accuracy of 82.78%.