Table 1.

Classification results

Training set (horizontal)	TG	Waltz	AmylHex
Tested set (vertical)
	sliding window of length 5
TG	0.75 | 0.62	0.82 | 0.21	0.77 | 0.42
Waltz	0.62 | 0.60	0.69 | 0.60	0.59 | 0.51
AmylHex	0.69 | 0.60	0.84 | 0.31	0.81 | 0.47
	sliding window of length 6
TG	0.76 | 0.57	0.77 | 0.30	0.78 | 0.44
Waltz	0.54 | 0.45	0.69 | 0.61	0.61 | 0.43
AmylHex	0.48 | 0.57	0.82 | 0.25	0.79 | 0.47

AUC ROC of the classification results with two window lengths. To test if a classification pattern is observable in the negative datasets, the training and testing procedures were also applied on negative datasets (POSITIVE | NEGATIVE); the positive datasets are in bold. Training dataset is defined horizontally; testing dataset – vertically. Random classification (or no pattern in a dataset) would obtain 0.5 and an ideal classifier 1.