Figure 8

Heat matrix for QDA cross validation error rate. Results for synthetic data. Results for synthetic data The figure displays the heat matrix of the 10-fold cross validation error rate of the QDA classification rule, obtained for all pairwise combinations of features belonging to the first position of the ranking of block matches. Four synthetic data containing the weak marginal/strong bivariate signals of scenarios 1, 2, 3 and 4, along with 198 noisy features, were generated. The signal represents 1% of the dimension of the feature space. The red color is a hot spot corresponding to a low error rate and a fancy interaction; meanwhile light yellow spots represent large error rates corresponding to pairs of noisy features.