Influence of secondary data sources and classifier combination on the performance for 10 training microarray data sets.

§secondary data source or combination rule with the best performance (W = weighting according to the 5-fold training AUC values; SW = weighting with scaled 5-fold training AUC values; EXH = exhaustive search among all possible combinations of 3 classifiers; NB = naïve Bayes; LR = logistic regression; LDA = linear discriminant analysis). *mean test AUC value under best training (standard deviation). °one-sided paired-sampled t-test for the comparison with respect to the baseline model.