TABLE 2.

Performance indices of different machine learning algorithms (MLP, GRNN, and ANFIS) for modeling and predicting shoot length, root length, number of nodes, number of shoots, and canopy surface area of Cannabis.

Model	Performance index	Shoot length		Shoot number		Node number		Root length		Canopy surface area
		Training	Testing	Training	Testing	Training	Testing	Training	Testing	Training	Testing
MLP	R2	0.972	0.954	0.625	0.421	0.717	0.390	0.938	0.900	0.953	0.928
	RMSE	4.929	6.927	0.396	0.632	0.702	1.202	15.112	15.956	277.487	340.538
	MBE	–0.090	1.673	0.009	–0.001	0.017	0.260	0.001	2.259	30.952	25.016
GRNN	R2	0.983	0.964	0.733	0.714	0.791	0.744	0.941	0.914	0.962	0.944
	RMSE	3.879	6.081	0.347	0.606	0.594	0.933	14.754	14.972	248.737	300.911
	MBE	0.001	1.540	0.001	0.012	–0.001	0.063	0.001	2.581	0.001	2.388
ANFIS	R2	0.770	0.590	0.647	0.501	0.767	0.549	0.781	0.589	0.733	0.644
	RMSE	17.538	23.327	0.407	0.557	0.650	0.942	41.881	39.007	1282.011	1282.697
	MBE	–4.549	–5.508	0.006	–0.065	–0.003	0.037	5.962	8.546	–26.525	–32.037

ANFIS, adaptive neuro-fuzzy inference system; GRNN, generalized regression neural network; MBE, mean bias error; MLP, multilayer perceptron; R², coefficient of determination; RMSE, root mean square error.