Table 1.

Discrimination results of Chrysanthemum varieties by different models using full wavelengths and optimal wavelengths.

Models	Full Wavelengths			Optimal Wavelengths
	Parameters 1	Training	Testing	Parameters	Training	Testing
SVM	(106, 10−5)	99.83%	94.02%	(107, 10−4)	98.26%	90.03%
LR	(L2, 100, liblinear)	99.34%	96.59%	(L2, 100, liblinear)	94.35%	85.75%
DCNN	(4, 32, 93)	99.98%	99.98%	(3, 32, 125)	98.45%	94.27%

¹ The parameters of the discriminant models. (c, g) for SVM, (pi, c’, optimize_algo) for LR, and (num_convs, num_first_kernels, epoch) for DCNN.