Table 3.

Performance criteria of DTs for peanut/weed classification

DT	Feature selection/reduction method	Train			Test
		Kappa	RMSE	Accuracy (%)	Kappa	RMSE	Accuracy (%)
J48	–*	0.88	0.2472	93.89	0.83	0.2791	91.67
	CFS	0.91	0.1942	95.56	0.88	0.1968	93.75
	PCA	0.63	0.4200	81.67	0.63	0.4210	81.48
REP	–*	0.84	0.2679	92.22	0.82	0.2888	90.91
	CFS	0.86	0.2557	92.78	0.83	0.2895	91.67
	PCA	0.47	0.4653	73.33	0.48	0.4986	74.074
RT	–*	0.70	0.3873	85.00	0.70	0.3849	85.18
	CFS	0.87	0.2582	93.33	0.85	0.2722	92.59
	PCA	0.46	0.5217	72.78	0.42	0.5401	70.83

Bolditalic value indicates the most accurate DT classifier

^* No feature selection applied (classification using all input features)