Predictive modeling of Persian walnut (Juglans regia L.) in vitro proliferation media using machine learning approaches: a comparative study of ANN, KNN and GEP models

Table 2 Evaluation of different developed models using various statistics for PR, CW, STN, and Vit of Persian walnut through in vitro proliferation

Walnut	Output	Model	Train Set ^*			Test Set
Walnut	Output	Model	RMSE	MAE	R²	RMSE	MAE	R²
Chandler	PR	MLR	2.840	1.986	0.264	3.313	2.270	0.412
		KNN (4) ^**	1.121	0.814	0.909	1.756	1.257	0.672
		MLPNN (9-40-1) ^***	1.093	0.796	0.913	1.658	1.186	0.695
		GEP	1.496	1.093	0.805	1.711	1.299	0.802
	CW	MLR	0.225	0.163	0.640	0.190	0.146	0.696
		KNN (4)	0.068	0.046	0.955	0.094	0.063	0.923
		MLPNN (9–15-1)	0.063	0.045	0.960	0.083	0.058	0.931
		GEP	0.113	0.086	0.886	0.124	0.105	0.844
	STN	MLR	7.628	5.787	0.315	7.827	6.614	0.241
		KNN (5)	3.324	2.432	0.867	3.529	2.767	0.855
		MLPNN (9–30-1)	2.777	1.994	0.917	2.660	2.026	0.915
		GEP	3.980	3.080	0.814	3.910	3.156	0.807
	Vit	MLR	6.189	4.930	0.435	6.909	5.538	0.434
		KNN (3)	0.992	0.735	0.986	1.232	0.960	0.974
		MLPNN (9–30-1)	0.993	0.699	0.988	1.184	0.938	0.975
		GEP	2.461	1.901	0.906	3.371	2.584	0.853
Rayen	PR	MLR	1.310	1.041	0.181	1.388	1.191	0.178
		KNN (4)	1.049	0.843	0.471	1.159	0.970	0.377
		MLPNN (9–12-1)	1.008	0.832	0.512	1.169	1.000	0.358
		GEP	1.039	0.858	0.459	1.178	0.992	0.428
	CW	MLR	0.400	0.319	0.286	0.460	0.382	0.276
		KNN (4)	0.069	0.050	0.949	0.078	0.057	0.929
		MLPNN (9–30-1)	0.067	0.049	0.952	0.078	0.056	0.930
		GEP	0.207	0.160	0.853	0.223	0.178	0.839
	STN	MLR	12.757	10.150	0.046	10.170	8.273	0.341
		KNN (5)	2.671	2.087	0.821	3.378	2.805	0.812
		MLPNN (9–30-1)	2.574	2.47	0.837	3.224	2.606	0.831
		GEP	4.916	3.822	0.858	5.238	4.242	0.808
	Vit	MLR	7.033	5.383	0.370	7.101	5.627	0.299
		KNN (4)	1.228	0.948	0.980	1.211	1.022	0.977
		MLPNN (9–20-1)	1.197	0.909	0.982	1.192	0.975	0.978
		GEP	2.849	2.235	0.899	2.937	2.452	0.891

ISSN: 1746-4811