Skip to main content

Table 1 Confusion matrix to evaluate accuracy, precision, and recall of the algorithm model

From: Constructing an automatic diagnosis and severity-classification model for acromegaly using facial photographs by deep learning

Predicted severity Actual numbers in the test dataset Total
Score 1 Score 2 Score 3
A Accuracy, precision, and recall of our algorithm model
 Score 1 32 1 1 34
 Score 2 4 85 2 91
 Score 3 7 7 98 112
 Total 43 93 101 237
 Precision 94.1% 93.4% 87.5%  
 Recall 74.4% 91.4% 97.0%  
 F1-Measure 0.831 0.924 0.920  
 Total prediction accuracy - - - 90.7%
B Accuracy, Precision, and Recall of Physicians
 Score 1 33 3 0 36
 Score 2 5 83 6 94
 Score 3 5 7 95 107
 Total 43 93 101 237
 Precision 91.7% 88.3% 88.8%  
 Recall 76.7% 89.3% 94.1%  
 F1-Measure 0.835 0.888 0.914  
 Total prediction accuracy - - - 89.0%