Skip to main content

Table 1 Performance evaluation of LGG and GBM subtypes classification

From: DeepAutoGlioma: a deep learning autoencoder-based multi-omics data integration and classification tools for glioma subtyping

 

Methods

Performance measures (Average of 10 fold cross-validation)

Accuracy [95% CI]

Precision [95% CI]

Recall [95% CI]

F1-score [95% CI]

FPR [95% CI]

Gmean [95% CI]

MCC [95% CI]

LGG

ANN

95.40%

92.50%

92.73%

92.45%

0.03

95.28%

0.89

[95.39–95.41]

[92.48–92.52]

[92.72–92.74]

[92.33–92.57]

[0.02–0.038]

[95.27–95.29]

[0.87–0.91]

CNN

98.03%

97.67%

96.96%

96.97%

0.01

97.99%

0.96

[98.02–98.038]

[97.66–97.679]

[96.95–96.97]

[96.96–96.98]

[0.005–0.015]

[97.98–97.998]

[0.95–0.97]

GBM

ANN

92.19%

88.05%

89.77%

87.75%

0.03

94.76%

0.9

[92.16–92.22]

[88.00- 88.10]

[89.73–89.81]

[87.70–87.80]

[0.02–0.04]

[94.74–94.78]

[0.86–0.94]

CNN

94.07%

90.40%

91.18%

90.25%

0.02

96.51%

0.93

[94.04–94.10]

[90.35–90.45]

[91.13–91.23]

[90.20–90.30]

[0.01–0.03]

[96.49–96.53]

[0.89–0.97]