Regression And Classification Prediction Of Blood Glucose Based On Factor Analysis Coupled With XGBoost Model

The number of chronically ill patients in China ranks first in the world,and diabetes and its related complications are an important part of it.Diabetes is currently unable to cure.For those who want to understand their blood sugar levels carefully,and those who only want to know roughly whether they have diabetes,this The different needs of the two populations,scientific and effective intervention,prevention and treatment,reducing the incidence and improving the quality of life of patients is necessary.Artificial intelligence(AI)can process and analyze massive medical and health data.This paper uses Python 3.0 software to do two aspects of work(1)We propose a model of XGBoost coupling factor analysis(FA-XGBoost)for predicting diabetes,in which factor analysis(FA)is used to reduce characteristic variables.Dimensions,the data set after the factor analysis dimension reduction is divided into a training set and a test set,the training set is used to learn the XGBoost model,and the test set is used to test the model effect.(2)In order to verify the effect of the model,some medical data provided by the Tianchi Precision Medical Competition were empirically analyzed from two aspects: regression prediction and classification prediction.In the regression prediction,the mean square error(MSE)and running time(t)of the FA-XGBoost regression model are 1.3800 and 1.3771 seconds respectively;on the classification problem,the correct rate,accuracy,recall rate of the FA-XGBoost classification model,The F1 values are 0.9543,0.9310,0.5197,and 0.6670,respectively.Finally,based on the same medical data,we compare the FA-XGBoost regression model and the classification model with the effects of the decision tree,random forest,GBDT,and XGBoost models.In the regression prediction,GBDT and FA-XGBoost performed best on MSE,while FA-XGBoost performed best in running time.In the classification problem,FA-XGBoost model showed performance slightly lower than the simple XGBoost model,but Best performance in accuracy,recall,and F1.Overall,the FA-XGBoost model is considered to perform well in both regression and classification predictions.The FA-XGBoost model is suitable for the prediction of blood glucose regression prediction,diabetes or no diabetes,and meets the needs of these two groups.For those who want to know their blood sugar levels,they can predict accurate blood sugar levels.People with diabetes can also predict whether they have diabetes by classification.