Table 1.

Accuracies of the SVM classification of the signals for each base for the signal-feature set that produced the most accurate overall classification of the bases.

	A accuracy (%)	G accuracy (%)	C accuracy (%)	T accuracy (%)
Equal beta training	88	83	71	55
Equal beta testing	70	51	55	35
Different beta training	95	80	62	60
Different beta testing	70	76	41	17

“Training” indicates that the cross validation was done on the same simulation.

“Testing” indicates that the cross validation was done with training on one simulation and testing on a different simulation.

The error in the accuracies is approximately ±5%.