Skip to main content

Table 2 Classification performance for different network architectures

From: Predicting enhancers with deep convolutional neural networks

Model

AUROC

AUPRC

Epoch Time

gkmSVM

0.887 (0.004)

0.899 (0.004)

6 h (total)

4conv2pool

0.910 (0.004)

0.915 (0.004)

272 s

4conv2pool4norm

0.916 (0.004)

0.917 (0.003)

376 s

4conv

0.896 (0.005)

0.897 (0.005)

325 s

6conv3pool

0.898 (0.005)

0.898 (0.006)

251 s

6conv3pool6norm

0.911 (0.006)

0.909 (0.005)

415 s

  1. The conventional gkmSVM is used as the baseline for comparison. For each model, we carried out 10-fold cross validation experiments. This table records the mean value of AUC values with standard error behind in the brackets
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us