Classification of patients with diabetes by analyzing the patterns of proteins present in saliva using machine learning

Modern equipment has allowed the mapping of biological information from a molecular point of view, which can be used for disease prediction and diagnosis. Besides, laboratory tests support clinical diagnosis and bring markers with complex patterns. These patterns can be used together with Machine Learning techniques to detect and automate patient diagnoses. These application tools can be used for the early diagnosis of diseases. This paper shows an application of supervised machine learning techniques to classify patients with diabetes using the protein markers in the saliva directly from the data provided by one piece of equipment deployed to measure molecular weights. Our database has 170 patients with 52 diabetics and 118 non-diabetic. We analyzed four widely used machine learning techniques, showing that the support vector machine achieved the best results concerning the accuracy, precision, recall, and F1-score.