Skip to main content

Table 4 Accuracies obtained by NN and classical machine learning methods on the three classification tasks

From: Assessment of deep learning and transfer learning for cancer prediction based on gene expression data

Classifier

Microarray

TCGA can.

TCGA type

XGBOOST

92.56 ± 0.29

99.03 ± 0.27

98.50 ± 0.14

LASSO

93.79 ± 0.35

97.70 ± 0.22

98,46 ± 0.14

RF

93.76 ± 0.29

98.41 ± 0.11

97,39 ± 0.21

SVMlin

93.81 ± 0.19

98.45 ± 0.09

98,70 ± 0.09

SVMrad

94.75 ± 0.25

98.66 ± 0.11

98,51 ± 0.09

NN

96.18** ± 0.18

99.09 ± 0.25

98.89** ± 0.18

  1. Bold highlights the methods with the best accuracy
  2. The symbol ** indicates that the accuracy of NN is significantly higher than the other methods (p value \(<0.01\) from the paired t-test)
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us