Fig. 3
From: EnsembleSplice: ensemble deep learning model for splice site prediction
Cross-Validation Ensemble model architecture. These are the architectural representation of each Ensemble model architecture and their individual base model architecture combination used in the cross-validation experiment. This contain a Ensemble ENS1 contains all DNN’s (DNN1, DNN2, DNN3, DNN4); b Ensemble ENS2 contains all the CNNs (CNN1, CNN2, CNN3, CNN4); c Ensemble ENS3 contains all the neural network models (DNN1, DNN2, DNN3, DNN4, CNN1, CNN2, CNN3, CNN4); d Ensemble ENS4 consists of CNN1, CNN2, CNN3, DNN1, DNN3; e Ensemble ENS5 consists of DNN1, DNN3, DNN4, CNN1, CNN2, CNN3; f Ensemble ENS6 consist of DNN1, DNN3, DNN4, CNN1, CNN2. We selected the Ensemble ENS2 from our cross-validation experiment