Skip to main content

Table 5 Wilcoxon paired signed-rank test was employed to compare the performance of DNN against the rest algorithms NB, kNN, RF and SVM across the datasets, with confidence intervals 99%

From: Deep-learning: investigating deep neural networks hyper-parameters and comparison of performance to shallow methods for modeling bioactivity data

Algorithms

Mean MCC diff.

SD of MCC diff.

p value

DNN-NB

0.149

0.061

4.768E−07

DNN-kNN

0.092

0.095

4.768E−07

DNN-SVM (linear)

0.052

0.031

3.2E−5

DNN-RF

0.021

0.016

8.6E−5

DNN-SVM (rbf)

0.009

0.012

5.075E−4

  1. Below are reported the means and standard deviations of the observed MCC for each algorithm and the p value
Back to article page

Journal of Cheminformatics

ISSN: 1758-2946

Contact us