Is possible to apply SVM with proc optmodel?

[carpdr]

you can... kernel-linear

results are not as good as kernel=polynom

Classification Table (Training Data)
Misclassification=1 Accuracy= 90.00
Observed	Predicted
	-1	1	Missing
-1	4.0000	1.0000	0
1	0	5.0000	0
Missing	0	0	0