A New Rule-Based Classification Method Using Shape-Based Properties of Spectral Curves

Table 3

Classification accuracies for plastic dataset.

Metrics	Rule-based method	SVM (a)	SVM (b)	SVM (c)
Overall accuracy	0.9694	0.9647	0.9641	0.9641
Precision	0.9604	0.9731	0.9403	0.9403
Sensitivity	0.9551	0.9272	0.9623	0.9623
False positive Rate	0.0077	0.0114	0.0076	0.0076
F1-score	0.9564	0.9487	0.9503	0.9503
Kappa	0.8603	0.8384	0.8358	0.8359

(a) Solver: C-SVC; Classification type: one-versus-all; kernel: RBF; C-value: 1; threshold: 0.5. (b) Solver: C-SVC; classification type: one-versus-all; kernel: linear; C-value: 1; threshold: 0.3. (c) Solver: C-SVC; classification type: one-versus-all; kernel: polynomial; C-value: 1; threshold: 0.3.