Diabetes mellitus(DM)is a metabolic disease,which is characterized by hyperglycemia.Long-term hyperglycemia may cause chronic damage to tissues.With the advent of big data and cloud computing era,artificial intelligence methods and ideas are widely used to analyze,and interpret diabetes-related data and,the method of designing highly precise and efficient algorithms have become a new research idea.For one of the specific characteristic of the nonlinearity and high dimensionality of diabetes clinical data,Support Vector Machine(SVM)is used in this paper,which can analyze the algorithm performs regression of blood glucose concentration.There are three research points: First,data preprocessing,including data transformation,data cleaning,etc.;second,feature selection,selecting the feature subset that minimizes cross-validation errors;third,constructing mixed kernel functions;and four,parameter optimization. |