Visualization of the classification performance for four classifiers and the set of features, selected at the 0.05 significance level.

<p>The observations along X axis are reordered according to their true class labels. For each observation red and green dots represent the estimated probabilities to belong to class 0 and 1 respectively. Dotted line separates observations from class 0 and class 1. As it is evident from the plot, the probability of observation to belong to a specific class is in agreement with its class label.