Journal of Translational Medicine

Table 3 Performance of the models based on different combinations of fingerprint descriptors and ML algorithms

From: Discovery of novel JAK1 inhibitors through combining machine learning, structure-based pharmacophore modeling and bio-evaluation

	Training set		Test set
	Accuracy	Precision	Accuracy	Precision	Recall	F1-score	Mcc	AUC
DNN
ECFP4	0.9996	0.9983	0.9928	0.9764	0.9937	0.9850	0.9803	0.9935
RDK	0.9976	0.9976	0.9913	0.9743	0.9896	0.9819	0.9762	0.9911
MACCS	0.9978	0.9972	0.9818	0.9466	0.9791	0.9626	0.9508	0.9812
SVM
ECFP4	0.9957	0.9956	0.9918	0.9763	0.9896	0.9829	0.9775	0.9928
RDK	0.9994	0.9976	0.9905	0.9762	0.9844	0.9803	0.9741	0.9886
MACCS	0.9905	0.9637	0.9803	0.9314	0.9906	0.9601	0.9478	0.9842
KNN
ECFP4	0.9973	0.9948	0.9915	0.9874	0.9771	0.9822	0.9767	0.9913
RDK	0.9925	0.9744	0.9866	0.9575	0.9875	0.9723	0.9636	0.9871
MACCS	0.9855	0.9497	0.9696	0.9036	0.9771	0.9389	0.9200	0.9725
LR
ECFP4	0.9992	0.9972	0.9915	0.9773	0.9875	0.9824	0.9768	0.9925
RDK	0.9995	0.9979	0.9898	0.9742	0.9833	0.9787	0.9720	0.9880
MACCS	0.9805	0.9567	0.9711	0.9305	0.9499	0.9401	0.9212	0.9641
RF
ECFP4	0.9993	0.9976	0.9918	0.9894	0.9760	0.9827	0.9773	0.9905
RDK	0.9993	0.9972	0.9925	0.9884	0.9802	0.9843	0.9794	0.9930
MACCS	0.9995	0.9979	0.9878	0.9820	0.9666	0.9743	0.9663	0.9809
DT
ECFP4	0.9745	0.9931	0.9674	0.9694	0.8916	0.9288	0.9091	0.9417
RDK	0.9696	0.9257	0.9686	0.9263	0.9437	0.9349	0.9143	0.9722
MACCS	0.9411	0.9127	0.9251	0.8697	0.8071	0.8372	0.7896	0.8848

Back to article page

ISSN: 1479-5876

Contact us

Submission enquiries: Access here and click Contact Us
General enquiries: info@biomedcentral.com