Diabetes Prediction Based on XGBoost Algorithm
Mingqi Li, Xiaoyang Fu, Dongdong Li
Abstract
Abstract Exploring important features of diabetes through analytical methods of data mining is able to predict and prevent diabetes. This paper proposes a diabetes prediction algorithm based on XGBoost algorithm with the numerical features being separated while some important features are extracted from the text features of experiment data. Experiment results show that accuracy of diabetes prediction based the improved XGBoost algorithm with features combination is 80.2%, which is feasible and effective method for diabetes prediction.
Topics & Concepts
Diabetes mellitusComputer scienceAlgorithmData miningArtificial intelligenceMachine learningMedicineEndocrinologyNutritional Studies and DietArtificial Intelligence in HealthcareCardiovascular Health and Risk Factors