Skip to main content

Table 2 Performance of PSO‐OA and literature similarity functions

From: Optimization and visualization of the edge weights in optimal assignment methods for virtual screening

 

AUC

BEDROC

awROCE5%

Target

PSO

OA

ECFP

FAP4D

PSO

OA

ECFP

FAP4D

PSO

OA

ECFP

4DFAP

ACE

0.947

0.728

0.815

0.882

0.678

0.459

0.420

0.521

15.6

11.9

11.8

11.6

AChE

0.906

0.713

0.795

0.762

0.704

0.445

0.788

0.521

13.0

5.0

11.9

8.6

CDK2

0.765

0.521

0.510

0.776

0.279

0.162

0.116

0.212

6.3

3.0

1.7

4.0

COX2

0.963

0.879

0.853

0.892

0.881

0.630

0.627

0.810

15.5

8.4

7.3

11.7

EGFr

0.738

0.670

0.756

0.990

0.593

0.364

0.743

0.918

8.5

5.5

12.3

17.8

FXa

0.865

0.440

0.626

0.652

0.420

0.065

0.046

0.094

8.5

1.8

1.2

2.9

HIVRT

0.690

0.519

0.630

0.575

0.134

0.221

0.316

0.252

3.3

3.0

3.9

2.3

InhA

0.946

0.542

0.747

0.668

0.777

0.486

0.628

0.611

16.7

8.3

9.7

8.7

P38

0.801

0.444

0.266

0.674

0.252

0.093

0.096

0.070

10.3

4.0

1.7

3.0

PDE5

0.720

0.499

0.424

0.667

0.308

0.228

0.208

0.351

4.5

5.5

2.1

4.5

PDGFrb

0.851

0.481

0.437

0.640

0.633

0.216

0.139

0.145

12.7

4.4

3.7

4.1

SRC

0.789

0.454

0.356

0.496

0.398

0.138

0.001

0.066

10.1

4.7

0.0

3.1

VEGFr2

0.853

0.354

0.391

0.684

0.329

0.121

0.063

0.169

8.6

2.6

0.8

3.5

avg.rank

1.23

3.54

3.15

2.08

1.54

3.08

3.08

2.31

1.30

2.77

3.23

2.69

  1. 1AUC, BEDROC, and awROCE5% performance of the PSO‐OA (PSO), OA, ECFP, and 4DFAP. Bold values indicate the best results on the corresponding data set and metric whereas italics indicate results that are statistically indistinguishable from the best result. The last row contains the average rank of the corresponding approach with respect to the metric and the other approaches.