Skip to main content

Table 2 Multi-modality integration simulation with MNIST dataset

From: Deep learning based feature-level integration of multi-omics data for breast cancer patients survival analysis

Modalities Random Erasing (S1) Gaussian Noise (S2)
ACC Precision Recall ACC Precision Recall
X1 0.942 ± 0.004 0.942 ± 0.004 0.942 ± 0.004 0.884 ± 0.003 0.886 ± 0.003 0.884 ± 0.003
X2 0.942 ± 0.003 0.943 ± 0.003 0.942 ± 0.003 0.879 ± 0.005 0.881 ± 0.005 0.879 ± 0.005
ConcatAE(X1 + X2) 0.962 ± 0.001 0.963 ± 0.001 0.962 ± 0.001 0.924 ± 0.001 0.925 ± 0.002 0.924 ± 0.001
CrossAE(X1 + X2) 0.962 ± 0.002 0.962 ± 0.002 0.962 ± 0.002 0.933 ± 0.002 0.933 ± 0.002 0.933 ± 0.002
\