Radiation Oncology

Table 3 Comparison of the classification performance for different models in the test dataset

From: Improving the performance of deep learning models in predicting and classifying gamma passing rates with discriminative features and a class balancing technique: a retrospective cohort study

	AUC			SEN			SPE
	MSE	SMSE	WMSE	MSE	SMSE	WMSE	MSE	SMSE	WMSE
FS1	0.78	0.77	0.79	0.45 (10/22)	0.55 (12/22)	0.55 (12/22)	0.90 (330/366)	0.89 (325/366)	0.87 (320/366)
FS2	0.78	0.79	0.83	0.45 (10/22)	0.55 (12/22)	0.59 (13/22)	0.90 (329/366)	0.89 (327/366)	0.88 (321/366)
FS3	0.80	0.82	0.86	0.50 (11/22)	0.59 (13/22)	0.64 (14/22)	0.90 (330/366)	0.90 (329/366)	0.89 (325/366)
FS4	0.83	0.86	0.87	0.59 (13/22)	0.64 (14/22)	0.68 (15/22)	0.91 (332/366)	0.89 (327/366)	0.89 (324/366)
FS5	0.88	0.88	0.92	0.64 (14/22)	0.68 (15/22)	0.77 (17/22)	0.91 (334/366)	0.90 (329/366)	0.89 (326/366)
FS6	0.86	0.86	0.91	0.64 (14/22)	0.68 (15/22)	0.73 (16/22)	0.90 (331/366)	0.90 (329/366)	0.90 (329/366)
FS7	0.85	0.82	0.87	0.59 (13/22)	0.59 (13/22)	0.68 (15/22)	0.90 (328/366)	0.89 (326/366)	0.89 (324/366)
FS8	0.81	0.81	0.84	0.50 (11/22)	0.55 (12/22)	0.59 (13/22)	0.91 (332/366)	0.89 (325/366)	0.88 (323/366)

FS = feature set; AUC = the area under the receiver operating characteristic curve; SEN = sensitivity; SPE = specificity; MSE = mean squared error; SMSE = MSE loss with re-sampling; WMSE = weighted MSE

Back to article page

ISSN: 1748-717X

Contact us

General enquiries: journalsubmissions@springernature.com