Machine Learning in the Field of Diabetes Prediction

Diabetes prevalence is a serious issue faced by all human beings. Tackling this issue can greatly help to reduce the financial burden on governments of medical expenditure to deal with diabetes patients and improve personal health by avoiding negative effects and complications brought by diabetes. To solve this problem, combining machine learning with the construction of diabetes prediction model has proved to be effective and is expected to develop rapidly in the future. This paper introduces five techniques commonly utilized in the construction of diabetes prediction model based on machine learning in detail, including Feature Ranking, AUC and C-statistics, Remote Framework for Diabetes Prediction, Gradient Boosting Algorithm and Model Evaluation Metrics. Based on these techniques, researchers can efficiently conduct relevant studies towards building diabetes prediction model with optimal machine learning algorithm selections. In the future, by further exploration on enhancing the combinations of prediction model with various machine learning algorithms and medical methodologies, the more accurate and effective ML-based diabetes prediction model can be built.