Skip to main content

Table 3 Performance comparison of prediction models

From: A novel approach to predicting the synergy of anti-cancer drug combinations using document-based feature extraction

Algorithms

FFNN

AE

XGB

ERT

LR

Ref.

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

(a) Performance comparison using ROC-AUC

 Baseline

0.912 ± 0.004

0.91 ± 0.005

0.914 ± 0.006

0.895 ± 0.007

0.885 ± 0.01

0.895 ± 0.005

0.843 ± 0.007

0.847 ± 0.006

 Ours

0.924 ± 0.001

0.915 ± 0.006

0.923 ± 0.003

0.92 ± 0.004

0.889 ± 0.003

0.892 ± 0.004

0.881 ± 0.004

0.854 ± 0.007

(b) Performance comparison using AUPR

 Baseline

0.402 ± 0.017

0.417 ± 0.032

0.408 ± 0.025

0.339 ± 0.026

0.349 ± 0.016

0.381 ± 0.026

0.24 ± 0.015

0.192 ± 0.011

 Ours

0.434 ± 0.014

0.427 ± 0.025

0.438 ± 0.008

0.424 ± 0.011

0.371 ± 0.013

0.381 ± 0.016

0.326 ± 0.012

0.196 ± 0.004

(c) Performance comparison using F1 score

 Baseline

0.359 ± 0.027

0.262 ± 0.045

0.325 ± 0.039

0.224 ± 0.05

0.227 ± 0.01

0.24 ± 0.019

0.272 ± 0.018

0.059 ± 0.01

 Ours

0.392 ± 0.017

0.313 ± 0.052

0.296 ± 0.021

0.428 ± 0.028

0.271 ± 0.016

0.259 ± 0.02

0.263 ± 0.015

0.064 ± 0.013