Skip to main content
. 2023 May 9;13:7544. doi: 10.1038/s41598-023-34303-8

Table 3.

Ablations on the intermediate model.

Model AUC Accuracy Specificity Sensitivity
Final component architecture
 FC layer 0.9* [0.82–0.95] 0.89 0.9 0.8
 1DCNN 0.91* [0.82–0.95] 0.89 0.89 0.9
 XGBoost 0.92* [0.84–0.97] 0.88 0.88 0.9
Dimensionality reduction
 w/o 0.92* [0.87–0.97] 0.85 0.84 0.9
 Sparse PCA 0.94 [0.88–0.98] 0.89 0.89 0.9
 NMF 0.94 [0.88–0.98] 0.85 0.84 0.9
 Modified LLE 0.96 [0.92–1.0] 0.9 0.92 0.8
 Spectral embedding 0.93* [0.89–0.97] 0.85 0.84 0.9
 BiAttention 0.95 [0.9–0.99] 0.9 0.9 0.9
Embedding dimension
 d = 1024 0.95 [0.9–0.98] 0.89 0.89 0.9
 d = 256 0.96 [0.93–1.0] 0.92 0.92 0.9
 Full 0.96[0.93-1.0] 0.93 0.94 0.9

Asterisks mark statistically significant difference compared to the full intermediate model architecture (Kolmogorov-Smirnov test with p0.05 ).

Best performing values are in [bold].