Table 4.

SVM classification performance. Support vector machine (SVM) classification performance for the experiment containing three treatments (control, mock-inoculated, and inoculated seedlings).

Comparison	Model	Data set	Accuracy	10-fold CV accuracy	Proportion correctly classified
					Control	Mock-inoculated	Inoculated
All groups	Second derivative	Training	0.822	0.708	0.825	0.786	0.854
		Testing	0.733	—	0.788	0.600	0.811
	VSURF	Training	0.830	0.664	0.775	0.810	0.899
		Testing	0.714	—	0.576	0.600	0.946
	Resampled	Training	0.866	0.644	0.813	0.905	0.876
		Testing	0.657	—	0.606	0.600	0.757
Mock vs. inoculated	Second derivative	Training	1.000	0.757	—	1.000	1.000
		Testing	0.806	—	—	0.829	0.784
	VSURF	Training	0.890	0.832	—	0.810	0.966
		Testing	0.861	—	—	0.829	0.892
	Resampled	Training	0.936	0.809	—	0.881	0.989
		Testing	0.847	—	—	0.800	0.892
Control vs. inoculated	Second derivative	Training	0.911	0.811	0.850	—	0.966
		Testing	0.886	—	0.848	—	0.919
	VSURF	Training	0.763	0.746	0.688	—	0.831
		Testing	0.643	—	0.485	—	0.784
	Resampled	Training	0.710	0.704	0.650	—	0.764
		Testing	0.643	—	0.545	—	0.730