Table 3.

5-Fold cross-validation performances of methods on Pan dataset

Method	Accuracy (%)	Precision (%)	Recall (%)	Specificity (%)	F1-Score (%)	MCC (%)
PIPR (2019)	98.26 ± 0.02	98.68 ± 0.04	97.40 ± 0.04	97.93 ± 0.03	98.04 ± 0.02	96.49 ± 0.03
FSNN-LGBM (2021)	99.50 ± 0.28	98.48 ± 0.12	99.39 ± 0.54	99.58 ± 0.10	99.43 ± 0.32	98.98 ± 0.57
Graph-BERT (2023)	99.02 ± 0.13	98.94 ± 0.88	99.15 ± 0.95	98.57 ± 1.19	99.04 ± 0.10	98.00 ± 0.28
Our xCAPT5	99.77 ± 0.02	99.75 ± 0.03	99.75 ± 0.02	99.80 ± 0.02	99.62 ± 0.06	99.55 ± 0.03

NA denotes that data is not available. Report with mean and standard deviation. The bold is the best performance in each metric