Skip to main content
. Author manuscript; available in PMC: 2018 Dec 7.
Published in final edited form as: SAR QSAR Environ Res. 2017 Apr 20;28(4):297–310. doi: 10.1080/1062936X.2017.1311941

Table 4:

How often a network predicted the LLNA potency class correctly, for 141 compounds as well as how often it over or under predicted the LLNA potency class.

Over and Under Prediction of LLNA Potency class

Test Set Correct LLNA Potency Class Over predicted Under predicted
Original Network 656961% 182115% 172015%
No TIMES-SS 495245% 232720% 283025%
Toolbox Alerts 545750% 222619% 242721%
Toolbox Alerts with Metabolism and Autoxidation 525749% 242721% 232621%
Toxtree Alerts 515546% 242921% 242821%
Toxtree Alerts with Metabolism and Autoxidation 515546% 242921% 252821%

Where Toolbox denotes the OECD QSAR Toolbox