Skip to main content

Advertisement

Table 13 Comparison of different model’s generalization performance

From: Prediction of 8-state protein secondary structures by a novel deep learning architecture

Model Accuracy(%)
CRRNN 71.4 ±0.2
Without ResNet 70.7 ±0.2
3-layer with BLSTM 70.2 ±0.2
Without local bolck 71.1 ±0.3
Without 1D one kernel CNN filter 71.5 ±0.2
With 2-layer BGRU 70.5 ±0.1
Unidirectional GRU 67.2
With 1-layer BGRU 69.5
CRRNN with 42dim features input 71.3 ±0.2