Skip to main content

Table 2 Model ablation studies with different experiment settings

From: DeepAEG: a model for predicting cancer drug response based on data enhancement and edge-collaborative update strategies

Trainning settings

PCC

Drug feature only

0.8507

Drug feature w/ muti-omics

0.9255

Drug feature w/ muti-omics w/ T

0.9268

Drug feature w/ muti-omics w/ T w/ Edge

0.9304

Double data enhancement w/ muti-omics w/ T w/ Edge

0.9318

Triple data enhancement w/ muti-omics w/ T w/ Edge

0.9333

  1. We show the contribution of the edge update module, transformer module and data enhancement features. Each module provides a PCC improvement of between 0.3 and 0.5% for the final results
  2. Muti-omics represents Multiomics characteristics of genes and T represents the transformer module. Edge represent the edge feature update module
  3. The best performance values obtained by the model are shown in bold
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us