. 2021 Dec 19;11(12):1904. doi: 10.3390/biom11121904

Table 2.

Accuracy measures after various numbers of training iterations for two different feature sets and three data sets.

Data Set	Training Iterations	Features	Accuracy Measures
Data Set	Training Iterations	Features	RMSE, Å²	MAE, Å²	MPE,%	MdPE, %	Δ₉₀, %	R ²	r
Training set	14,160	Full	12.6	7.5	1.46	1.00	3.21	0.987	0.993
	26,000	Full	11.7	7.1	1.40	0.96	3.05	0.988	0.994
	76,110	Full	9.8	6.2	1.23	0.87	2.68	0.992	0.996
	30,000	Reduced	11.4	7.0	1.38	0.95	3.03	0.989	0.994
Test set	14,160	Full	13.1	7.7	1.50	1.02	3.28	0.985	0.993
	26,000	Full	12.8	7.6	1.47	1.00	3.25	0.986	0.993
	76,110	Full	13.2	7.6	1.48	0.99	3.22	0.985	0.993
	30,000	Reduced	13.3	7.8	1.52	1.03	3.32	0.985	0.992
ProteomeTools test set	14,160	Full	14.9	9.4	1.84	1.28	4.02	0.983	0.991
	26,000	Full	14.6	9.3	1.82	1.28	3.95	0.983	0.992
	76,110	Full	15.3	9.5	1.86	1.27	4.06	0.982	0.991
	30,000	Reduced	15.4	9.7	1.90	1.32	4.17	0.982	0.991