Selection-fusion algorithm for missing value management: B is the set of all subsets, M is the missing value matrix as defined in Equation (1), A is the co-accordance binary matrix as defined in Equation (2), C is the subset similarity matrix as defined in Equation (3), D is the distance matrix as defined in Equation (5), CVI is the cluster validity index as defined in Equation (6), xj is the testing sample, φ̃i,xj is the fusion weight as defined in Equation (11), and ΓBB(x) is the classification result for the testing sample x as defined in Equation (9).