Fig. 5 | Scientific Reports

Fig. 5

From: Identifying and predicting gait stability metrics in people with stroke in uneven-surface walking using machine learning

Fig. 5

Partial dependence plot analysis and SHAP partial dependence plot analysis for gait parameters on uneven surface. Partial dependence plots (PDPs) and SHAP dependence plots for key even-surface gait features predicting uneven-surface gait outcomes using the Random Forest model. Each row corresponds to a target variable: (a) Gait speed, (b) RMS_VT, (c) SampEn_AP, and (d) HR_AP. For each variable, the left panel shows the PDP, and the right panel shows the SHAP dependence plot with individual data points colored by feature value. Plots are presented for the most influential same-name feature (top row) and an additional relevant predictor (bottom row) identified through SHAP analysis. RMS, Root mean square; VT, Vertical; Samp_En, Sample entropy; AP, Anterior-posterior; HR, Harmonic ratio.

Back to article page