Skip to main content
Fig. 2 | BMC Bioinformatics

Fig. 2

From: MATHLA: a robust framework for HLA-peptide binding prediction integrating bidirectional LSTM and multiple head attention mechanism

Fig. 2

Model evaluation by the fivefold cross-validation test. a Fivefold cross-validation test was repeated for 10 times. The mean and standard deviation of AUC scores were shown. b In the fivefold cross-validation test, data in each fold were stratified by peptide length. The AUC score for each length was then calculated. The mean and standard deviation of AUC scores of 10 repeated tests were shown

Back to article page