Egyptian Journal of Radiology and Nuclear Medicine

Table 2 Evaluation of model performance in detecting “Normal” chest radiographs

From: Classification of chest radiographs using general purpose cloud-based automated machine learning: pilot study

	Normal	Not normal	Total
Classified as “Normal”	185 (true positive)	76 (false positive)	261
Not classified as “Normal”	120 (false negative)	256 (true negative)	376
Total	305	332
Statistics	Value	95% confidence interval
Sensitivity	60.66%	54.93 to 66.17%
Specificity	77.11%	72.21 to 81.52%
Positive likelihood ratio	2.65	2.13 to 3.29
Negative likelihood ratio	0.51	0.44 to 0.59
Positive predictive value	70.88%	66.21 to 75.15%
Negative predictive value	68.09%	64.71 to 71.28%
Accuracy	69.23%	65.48 to 72.80%

Back to article page