Using Machine Learning to Unravel the Value of Radiographic Features for the Classification of Bone Tumors

Table 4

The performance of random forest models.

	AUC	Sensitivity	Specificity	Accuracy	Overall accuracy	Micro AUC
Binary model	0.97	93.33%	95.61%	94.71%
Tertiary model
Benign	0.95	86.18%	83.62%	84.94%	82.77%	0.94
Malignant	0.98	90.77%	92.53%	92.05%
Intermediate	0.89	58.82%	93.62%	86.19%

Note: AUC: area under the receiver operating characteristic curve. All the results were obtained in the test set of two models.