Research Paper Volume 15, Issue 10 pp 4465—4480
Identification and validation of diagnostic signature genes in non-obstructive azoospermia by machine learning
Figure 4. LightGBM model and model performance. (A) Receiver operating curves (ROC) for signature genes in the training dataset. (B) Precision-Recall curves for signature genes in the training dataset. (C) Receiver operating curves (ROC) for signature genes in the external validation dataset. (D) Precision-Recall curves for signature genes in the external validation dataset. (P < 0.05).