Skip to main content

Advertisement

Table 4 Mean performance of similarity methods across MUV data sets.

From: Large scale study of multiple-molecule queries

Method AUC F1 BEDROC
MIN-RANK 0.731133 ± 0.030578 0.149965 ± 0.023025 0.345171 ± 0.042642
MAX-RANK 0.509469 ± 0.020590 0.017739 ± 0.004382 0.061569 ± 0.010419
SUM-RANK 0.598784 ± 0.030562 0.021604 ± 0.005490 0.104799 ± 0.022261
MAX-SIM 0.714848 ± 0.028352 0.156955 ± 0.025644 0.312150 ± 0.041033
MIN-SIM 0.533202 ± 0.025204 0.020921 ± 0.004781 0.070374 ± 0.008572
SUM-SIM 0.617073 ± 0.034809 0.052993 ± 0.021674 0.153437 ± 0.040308
NUMDEN-SIM 0.644467 ± 0.032684 0.061232 ± 0.022654 0.177026 ± 0.040264
BAYES 0.642907 ± 0.031377 0.041962 ± 0.011625 0.176723 ± 0.039162
BKD 0.784118 ± 0.025509 0.145667 ± 0.027758 0.354250 ± 0.044227
ETD 0.785944 ± 0.025833 0.141997 ± 0.026715 0.356733 ± 0.043394
TPD 0.775774 ± 0.025289 0.152530 ± 0.023448 0.352975 ± 0.041960
SUM-EH 0.679485 ± 0.030601 0.100943 ± 0.025459 0.230528 ± 0.046206
SUM-ET 0.733893 ± 0.027988 0.155680 ± 0.026528 0.324622 ± 0.042935
SUM-TP 0.729849 ± 0.028113 0.157239 ± 0.026507 0.323068 ± 0.042618
  1. The mean performance of similarity methods across the 17 MUV data sets with their corresponding backgrounds. A confidence interval is provided with each measurement. The best performance in each column is listed in bold face, and all performances statistically indistinguishable (with a t-test yielding a p-value > 0.05) are listed in italics.