Figure 2

Meta-classification. The unlabeled sample is first tested against all the binary classifiers and Figure 2(a) shows the predictions obtained after such a test. Next, its bit sequence representation is constructed [Figure 2(b)]. The XOR operation is performed between this bit sequence and each of the sequences from the meta-data set [Figures 2(c) & 2(d)]. The distance between samples is the number of 1-bits in the XOR operation result. Since the distance of the unlabeled sample from sample 1 is less than that from sample 2, the unlabeled sample is assigned the label of sample 1, i.e. it is assigned to the class a.