Skip to main content

Table 5 Effect of hash size on predictive performance with Scikit RF, Fest and Scikit SVM

From: Evaluating parameters for ligand-based modeling with random forest on sparse data sets

Method Radius Hash size Mean SD
    ROC AUC  
Scikit RF 1 128 0.849175 0.0023297
Scikit RF 1 256 0.854675 0.00229456
Scikit RF 1 512 0.857525 0.00262345
Scikit RF 1 1024 0.85965 0.00294236
Scikit RF 1 2096 0.85945 0.00357561
Scikit RF 1 4096 0.85945 0.00348676
Scikit RF 1 Unhashed 0.879925 0.00144135
Scikit RF 2 128 0.843675 0.00222542
Scikit RF 2 256 0.853675 0.0033908
Scikit RF 2 512 0.859175 0.00176494
Scikit RF 2 1024 0.863025 0.0014151
Scikit RF 2 2096 0.8633 0.00241454
Scikit RF 2 4096 0.8644 0.00129808
Scikit RF 2 Unhashed 0.87795 0.00214126
Scikit RF 3 128 0.834875 0.00240936
Scikit RF 3 256 0.8472 0.00245917
Scikit RF 3 512 0.85435 0.00269165
Scikit RF 3 1024 0.858525 0.00269258
Scikit RF 3 2096 0.8603 0.0031249
Scikit RF 3 4096 0.862875 0.00200624
Scikit RF 3 Unhashed 0.876575 0.00171391
FEST 1 128 0.849275 0.00152151
FEST 1 256 0.856275 0.00226661
FEST 1 512 0.85865 0.00271616
FEST 1 1024 0.8594 0.00257633
FEST 1 2096 0.860975 0.00175784
FEST 1 4096 0.861525 0.00239008
FEST 1 Unhashed 0.879325 0.00281158
FEST 2 128 0.844275 0.00282975
FEST 2 256 0.8531 0.00292104
FEST 2 512 0.85865 0.00290086
FEST 2 1024 0.8639 0.00189143
FEST 2 2096 0.864225 0.00191703
FEST 2 4096 0.86575 0.00242178
FEST 2 Unhashed 0.87755 0.00140624
FEST 3 128 0.83445 0.00234254
FEST 3 256 0.84695 0.00355633
FEST 3 512 0.8536 0.00216102
FEST 3 1024 0.858325 0.00195704
FEST 3 2096 0.86105 0.00356686
FEST 3 4096 0.863875 0.00230326
FEST 3 Unhashed 0.874625 0.00217658
Scikit SVM 1 128 0.845025 0.00236326
Scikit SVM 1 256 0.854125 0.00246475
Scikit SVM 1 512 0.85575 0.00195512
Scikit SVM 1 1024 0.857875 0.0023516
Scikit SVM 1 2096 0.858425 0.00126293
Scikit SVM 1 4096 0.85875 0.00224666
Scikit SVM 1 Unhashed 0.875175 0.00193197
Scikit SVM 2 128 0.83875 0.00233024
Scikit SVM 2 256 0.848125 0.00232433
Scikit SVM 2 512 0.85155 0.00185876
Scikit SVM 2 1024 0.856875 0.00197927
Scikit SVM 2 2096 0.860825 0.00186615
Scikit SVM 2 4096 0.862725 0.00172699
Scikit SVM 2 Unhashed 0.87635 0.00306716
Scikit SVM 3 128 0.8283 0.00287359
Scikit SVM 3 256 0.838625 0.00278298
Scikit SVM 3 512 0.843375 0.00289093
Scikit SVM 3 1024 0.849425 0.00141067
Scikit SVM 3 2096 0.85605 0.00269629
Scikit SVM 3 4096 0.858475 0.00194936
Scikit SVM 3 Unhashed 0.8784 0.0017713
  1. Each data point is a mean of the four data sets. These values are plotted as Fig. 2 in the publication