Figure 3 | Scientific Reports

Figure 3

From: Machine learning for post-acute pancreatitis diabetes mellitus prediction and personalized treatment recommendations

Figure 3

Model interpretation. We have used two methods of model interpretation. (A) Nomogram. The trend and magnitude of the effect of the nine core factors on the prediction of positive events can be observed in the figure. Admission glucose, BMI > 28, Age, NAFLD, UA and Smoking are the risk factors for PPDM-A. In contrast, Cardiovascular disease, ALT, HDL-C < 1.03 are negative predicted factors. (B) Sharpley value was used to explain the effect of the model on prediction. HDL-C < 1.03, BMI > 28 and Admission Glucose were the main factors affecting prediction. BMI > 28, Cardiovascular disease, HDL-C < 1.03 and Smoking are logistic variables, with 0 being FALSE and 1 being TRUE.

Back to article page