Table 2.

Performance comparison of proposed ADH-PPI predictor with 12 existing PPI predictors on benchmark S. cerevisiae dataset, where results of existing PPI predictors are taken from (Yu et al., 2021) paper

Method	ACC (%)	Recall (%)	Precision (%)	MCC
ACC+SVM (Guo et al., 2008a)	0.8933 ± 2.67	0.8993 ± 3.68	0.8887 ± 6.16	N/A
Code4+KNN (Guo et al., 2008a)	0.8615 ± 1.17	0.8103 ± 1.74	0.9024 ± 1.34	N/A
MCD+SVM (You et al., 2014)	0.9136 ± 0.36	0.9067 ± 0.69	0.9194 ± 0.62	0.8421 ± 0.0059
MLD+RF (You et al., 2015a)	0.9472 ± 0.43	0.9434 ± 0.49	0.9891 ± 0.33	0.8599 ± 0.0089
PR-LPQ+RF (You et al., 2015b)	0.9392 ± 0.36	0.9110 ± 0.31	0.9645 ± 0.45	0.8856 ± 0.0063
MIMI + NMBAC+ RF (Ding et al., 2016)	0.9501 ± 0.46	0.9267 ± 0.50	0.9716 ± 0.55	0.9010 ± 0.0092
LRA+RF (You et al., 2017)	0.9414 ± 1.8	0.9122 ± 1.6	0.9710 ± 2.1	0.8896 ± 0.026
DeepPPI (Du et al., 2017)	0.9443 ± 0.30	0.9206 ± 0.36	0.9665 ± 0.59	0.8897 ± 0.0062
ippi-esml (Jia et al., 2015)	0.9515 ± 0.25	0.9221 ± 0.36	0.9797 ± 0.60	0.9045 ± 0.0053
WSRC (Kong et al., 2020)	0.8673	0.8993	NA	0.7693
DeepFE-PPI (Yao et al., 2019)	0.9478	0.9299	0.9645	0.8962
GcForest-PPI (Yu et al., 2021)	0.9544	0.9272	0.9805	0.9102
ADH-PPI	0.9573	0.9394	0.9575	0.9144