Eq. (1) defines the distribution of a relation strength for irrelevant data entries φ. Eq. (2) defines an relation strength from the clsuter k to cluster l for relevant data entries. λi, i = 1, 2, . . . , N in Eq. (3) denotes the probability of a relevancy flag variable ri being 1. ri = {0, 1} in Eq. (4) indicates whether the object i is relevant or not. zi = k ∈ {1, 2, . . . } indicates the cl...