. 2023 Dec 13;12(24):4151. doi: 10.3390/plants12244151

Table 4.

Machine learning algorithm goodness-of-fit criteria for predicting callus induction (CI), embryogenic callus induction (EC), and regeneration efficiency (RE).

Traits	ML Criteria	SVM		RF		XGBoost		KNN		GP
Traits	ML Criteria	Train	Test	Train	Test	Train	Test	Train	Test	Train	Test
CI ¹	R²	0.281	0.098	0.402	0.244	0.551	0.515	0.076	0.068	0.539	0.443
	MSE	10.462	16.620	9.545	15.217	8.273	12.190	11.859	16.890	8.379	13.055
	MAPE	10.438	21.635	10.267	19.753	8.607	16.248	12.025	21.691	9.028	17.425
	MAD	7.761	12.118	7.876	11.287	6.677	9.523	9.298	12.172	7.006	10.332
EC	R²	0.383	0.574	0.436	0.719	0.648	0.393	0.144	0.432	0.595	0.706
	MSE	13.324	9.326	12.739	7.577	10.069	11.130	15.694	10.768	10.798	7.743
	MAPE	15.835	9.899	20.133	9.329	14.870	16.650	24.518	15.233	15.825	10.389
	MAD	8.665	6.923	10.472	6.165	8.334	9.983	12.547	9.169	8.522	6.916
RE	R²	0.526	0.505	0.502	0.422	0.671	0.461	0.145	0.236	0.659	0.525
	MSE	0.185	0.173	0.190	0.186	0.155	0.180	0.249	0.214	0.157	0.169
	MAPE	37.895	55.642	56.053	59.253	34.188	54.184	72.593	69.556	36.623	48.638
	MAD	0.131	0.154	0.161	0.167	0.121	0.159	0.201	0.191	0.126	0.138

¹ CI: callus induction (%); EC: embryogenic callus induction (%); and RE: regeneration efficiency (number).