Table 7.

Random forest algorithm results for temporal layer stacked image.

Random forest algorithm results for temporal layer stacked image
Training 70							Testing 30
Maximum depth		Fields_ training	Forest_ training	Shrubs_ training	Urban_ training	Overall. accuracy	Fields_ testing	Forest_ testing	Shrubs_ testing	Urban_ testing	Overall. accuracy
5	Overall accuracy = 94.07%, kappa coefficient = 0.903						Overall accuracy = 94.79%, kappa coefficient = 0.916
	Unclassified	0	0	0	0		0	0	0	0
	Fields	16,540	182	144	234	89.98%	7079	98	92	107	91.06%
	Forest	622	32,050	326	46	96.22%	210	13,812	116	7	97.04%
	Shrubs	761	1076	3667	0	88.40%	276	323	1458	0	87.15%
	Urban	459	0	11	9071	97.01%	209	0	7	3974	97.21%
10	Overall accuracy = 97.49%, kappa coefficient = 0.961						Overall accuracy = 96.33%, kappa coefficient = 0.942
	Unclassified	0	0	0	0		0	0	0	0
	Fields	16,849	58	65	128	95.53%	7140	81	59	96	94.37%
	Forest	298	32,536	159	51	98.7%	121	13,901	111	12	97.83%
	Shrubs	260	377	4867	0	95.68%	171	221	1665	0	90.49%
	Urban	231	1	5	9304	98.11%	134	7	5	4044	97.40%
20	Overall accuracy = 99.79%, kappa coefficient = 0.996						Overall accuracy = 96.98%, kappa coefficient = 0.954
	Unclassified	0	0	0	0		0	0	0	0
	Fields	17,063	0	6	31	99.56%	7158	91	33	94	96.18%
	Forest	36	32,986	15	7	100%	69	13,977	60	12	97.74%
	Shrubs	6	0	5498	0	99.62%	114	227	1716	0	94.70%
	Urban	33	0	0	9508	99.60%	101	5	3	4081	97.47%