BMC Endocrine Disorders

Table 2 Summary of model performance of C-index, Accuracy, Sensitivity, Specificity, Brier-score

From: Comparison of the cox regression to machine learning in predicting the survival of anaplastic thyroid carcinoma

	Trainingset	Testset
	C-index	C-index	Accuracy	Sensitivity	Specificity	Brier score
Classification Models
OS 6 month
Logistic	0.779	0.790	0.720	0.757	0.724	0.170
Random Forests	0.834	0.728	0.691	0.737	0.653	0.181
XGBoost	0.808	0.763	0.703	0.696	0.737	0.169
AdaBoost	0.788	0.788	0.699	0.739	0.726	0.238
OS 12 month
Logistic	0.794	0.811	0.728	0.829	0.677	0.117
Random Forests	0.886	0.736	0.679	0.756	0.659	0.131
XGBoost	0.819	0.767	0.682	0.660	0.778	0.126
AdaBoost	0.803	0.786	0.696	0.784	0.691	0.192
CSS 6 month
Logistic	0.775	0.775	0.700	0.770	0.679	0.179
Random Forests	0.857	0.700	0.648	0.672	0.689	0.204
XGBoost	0.794	0.743	0.679	0.724	0.677	0.185
AdaBoost	0.774	0.779	0.709	0.772	0.685	0.238
CSS 12 month
Logistic	0.760	0.768	0.721	0.735	0.704	0.127
Random Forests	0.910	0.662	0.638	0.646	0.639	0.174
XGBoost	0.790	0.711	0.688	0.705	0.653	0.157
AdaBoost	0.775	0.761	0.700	0.680	0.758	0.244
Time-to-Event Models
OS
Cox	0.709	0.713
Survival Tree	0.667	0.630
Survival SVM	0.700	0.651
Random Survival Forests	0.668	0.630
XGBoost	0.720	0.657
DeepSurv	0.945	0.658
CSS
Cox	0.710	0.712
SurvivalTree	0.670	0.628
Survival SVM	0.701	0.644
Random Survival Forests	0.670	0.628
XGBoost	0.709	0.662
DeepSurv	0.834	0.676

Abbreviations: C-index: concordance index; XGBoost: Extreme Gradient Boosting; AdaBoost: Adaptive Boosting; SVM: Support Vector Machine; OS: overall survival; CSS: cancer-specific survival

Back to article page

ISSN: 1472-6823

Contact us

General enquiries: journalsubmissions@springernature.com