Skip to main content
Fig. 5 | BMC Medical Informatics and Decision Making

Fig. 5

From: Classification of painful or painless diabetic peripheral neuropathy and identification of the most powerful predictors using machine learning models in large cross-sectional cohorts

Fig. 5

Partial Dependence (PD) plots for the Adaptive Regression Splines model aggregated across imputations. For panels A–D: Y-axis show the average marginal effect on the prediction given a certain value of the respective variable/feature. X-axis show the feature's values and the rug of marks under the x-axis indicate the density of data distribution across the feature’s full range. Black lines with yellow outline represent the aggregated PD. Blue lines show the fitted smoothed LOESS curve and highlight the PD trend. Selected independent variables had a scaled importance of > 10 and are arranged by decreasing importance value

Back to article page