Table 1.

Performance statistics of aggregation models on training and test sets and by Monte-Carlo simulations

Set/method of validation	Number of antibodies	Overall accuracy	Accuracy in Low	Accuracy in High	Specificity in Low	Specificity in High
The κ model
training set	256	92%	89%	95%	95%	88%
MC-whole set	341	72%	75%	68%	75%	69%
test set	85	76%	78%	75%	81%	71%
The ‘κ and λ’ model
training set	355	91%	88%	95%	96%	84%
MC-whole set	472	68%	67%	71%	76%	60%
test set	117	70%	75%	64%	72%	68%