Fig. 5 | Scientific Reports

Fig. 5

From: Machine learning models to predict osteoporosis in patients with chronic kidney disease stage 3–5 and end-stage kidney disease

Fig. 5

Calibration curve (A) and decision curve analysis (B) for five selected models in osteoporosis prediction. For the calibration curve, the perfectly calibrated model is represented by the dashed diagonal. Deviations from this line indicate calibration issues, with models that either overestimate or underestimate the probabilities. For the decision curve analysis, the dashed gray line represents the “ALL” strategy, which assumes that every patient has osteoporosis, while the black line represents the “NONE” strategy, which assumes that no patient has osteoporosis. The x-axis represents the threshold probability, and the y-axis represents the net benefit. The models closer to the top of the graph provide better clinical utility across different threshold probabilities.

Back to article page