Skip to main content
Fig. 3 | Journal of Translational Medicine

Fig. 3

From: Prediction of 3-year risk of diabetic kidney disease using machine learning based on electronic medical records

Fig. 3

a Importance matrix plot of the LightGBM model, depicting the importance of each variable for predicting 3-year DKD risk in patients with T2DM and normo-albuminuria. b SHAP summary plot of the top 8 clinical features of the LightGBM model. There is one dot per patient per feature colored according to an attribution value, where red represents a higher value and blue represents a lower value. Hcy, homocysteine; BMI, body mass index; ALB, serum albumin; eGFR, estimated glomerular filtration rate; LDL, low-density lipoprotein

Back to article page