Research Paper Volume 13, Issue 21 pp 24219—24235

Construction and external validation of a 5-gene random forest model to diagnose non-obstructive azoospermia based on the single-cell RNA sequencing of testicular tissue

Validation of the diagnostic efficacy of the random forest model. (A, B) The ROC (A) and confusion matrix (B) of the predictive model in the training dataset. (C, D) The ROC (C) and confusion matrix (D) of the predictive model in the external validation dataset. ROC, receiver operating curve. AUC, area under curve; NOA, non-obstructive azoospermia; OA, obstructive azoospermia.

Figure 5. Validation of the diagnostic efficacy of the random forest model. (A, B) The ROC (A) and confusion matrix (B) of the predictive model in the training dataset. (C, D) The ROC (C) and confusion matrix (D) of the predictive model in the external validation dataset. ROC, receiver operating curve. AUC, area under curve; NOA, non-obstructive azoospermia; OA, obstructive azoospermia.