Classifying forest inventory data into species-based forest community types at broad extents: exploring tradeoffs among supervised and unsupervised approaches

Forest Ecosystems

Table 3 Results from indicator species analysis, species dominance, and random forest models for all classifications

Region	Classification	Indicator species metrics				Species dominance	Random forest metrics
Region	Classification	Sum of indicator vals.^a	Avg. of p-values	Num. significant indicator vals.^a	Prop. classes without significant indicator vals.	Mean of largest SDI (std. dev.)	Kappa	Avg. out-of-bag error
U.S.	FIA	1.01	0.01	6.00	0	0.47 (0.17)	0.59	35.74%
	Empirical	1.21	0.05	3.76	0	0.61 (0.21)	0.74	22.80%
East	USNVC	0.60	0.07	3.32	0	0.36 (0.16)	0.55	40.25%
	FIA	0.87	0.08	3.24	0.12	0.42 (0.17)	0.48	38.55%
	Empirical	0.88	0.19	1.35	0.24	0.59 (0.24)	0.64	24.90%

^aThe sum of indicator values and the number of significant indicator values reported here have been standardized by the number of classes in each classificationBold indicates the best value for each metric in each region