Machine learning and deep learning techniques in diabetes prediction

Under the joint influence of environment and genes, diabetes mellitus has now become a growing issue with a series of complications, which leads to a low quality of life. Patients with developing diabetes may suffer from limited diets, scheduled medication, physical pain, and mental torment. Considering the high morbidity, it is of great significance to predict whether a person has diabetes and to take action to alleviate the disease effectively. So far, there are many studies focusing on diabetes prediction with high efficiency and accuracy by introducing specific data mining techniques and algorithms, such as Machine Learning and Deep Learning. The practice of Machine Learning techniques including Decision Tree, XG boost, and Random Forest in diabetes prediction, has been illustrated by different authors and compared accordingly. Artificial Neural Network, one of the Deep Learning methods, being used to predict diabetes patients, was put forward by various researchers with the comparison of the accuracy. In this article, the factors that influence the prediction accuracy are discussed. The practical application of these methods is discussed as well, with the aim of obtaining a hi