Prediction of donor splice sites using random forest with a new sequence encoding approach

Table 5 Standard errors of different performance metrics for Bagging, Boosting, Logistic regression, kNN and Naïve Bayes classifiers for all the three encoding procedures under both balanced and imbalanced situations

EP	MD	Balanced							Imbalanced
EP	MD	TPR	TNR	F (α = 1)	F (β = 2)	G-mean	WA	MCC	TPR	TNR	F (α = 1)	F (β = 2)	G-mean	WA	MCC
P-1	BG	0.0201	0.0178	0.0114	0.0156	0.0113	0.0113	0.0226	0.0234	0.0036	0.0409	0.0282	0.0474	0.0108	0.0334
	BS	0.0146	0.0149	0.0111	0.0125	0.0113	0.0112	0.0224	0.0177	0.3156	0.0334	0.0218	0.0715	0.1652	0.0504
	LG	0.0569	0.0740	0.0601	0.0575	0.0624	0.0621	0.1238	0.0065	0.0056	0.0121	0.0076	0.0313	0.0045	0.0267
	NB	0.0630	0.0826	0.0560	0.0571	0.0573	0.0577	0.1169	0.0357	0.1043	0.0500	0.0390	0.0439	0.0549	0.1579
	KN	0.1502	0.1279	0.1386	0.1454	0.1364	0.1354	0.2701	0.0221	0.3023	0.0389	0.0267	0.0765	0.1595	0.0799
P-2	BG	0.0201	0.0272	0.0192	0.0188	0.0201	0.0200	0.0397	0.0261	0.0060	0.0429	0.0310	0.0421	0.0130	0.0364
	BS	0.0207	0.0179	0.0161	0.0184	0.0163	0.0163	0.0327	0.0273	0.0033	0.0456	0.0325	0.0461	0.0134	0.0358
	LG	0.0688	0.0799	0.0617	0.0644	0.0630	0.0632	0.1272	0.0182	0.0148	0.0290	0.0214	0.0273	0.0107	0.0410
	NB	0.0546	0.0629	0.0421	0.0472	0.0405	0.0407	0.0824	0.0316	0.0733	0.0487	0.0363	0.0436	0.0426	0.1342
	KN	0.0925	0.0811	0.0362	0.0681	0.0266	0.0280	0.0598	0.0235	0.0044	0.0337	0.0269	0.0282	0.0117	0.0270
P-3	BG	0.0156	0.0186	0.0117	0.0130	0.0120	0.0119	0.0237	0.0185	0.0052	0.0291	0.0217	0.0267	0.0089	0.0235
	BS	0.0121	0.0178	0.0102	0.0102	0.0108	0.0107	0.0210	0.0194	0.0039	0.0324	0.0231	0.0323	0.0095	0.0256
	LG	0.0406	0.0586	0.0376	0.0377	0.0409	0.0402	0.0795	0.0210	0.0116	0.0334	0.0247	0.0303	0.0132	0.0440
	NB	0.0380	0.0689	0.0330	0.0323	0.0372	0.0368	0.0735	0.0254	0.0434	0.0397	0.0295	0.0333	0.0286	0.0913
	KN	0.1017	0.0829	0.0629	0.0842	0.0566	0.0544	0.1076	0.0292	0.0078	0.0504	0.0352	0.0629	0.0116	0.0334

MD methods, EP encoding procedures, BG bagging, BS boosting, LG logistic regression, NB naïve bayes, KN K nearest neighbor

ISSN: 1756-0381