Figure 1
From: Enhanced protein fold recognition through a novel data integration approach
Synthetic data verification. Synthetic data verification of MKLdiv: (a) depiction of the three-circle dataset; (b) relative change of objective values of MKLdiv-dc versus iteration number of CCCP; (c) relative change of objective values of MKLdiv-conv versus iteration number of projected gradient descent (PGD) method; (d) relative change of objective values of subproblem (13) by SILP (dish-line) and PGD methods; (e) kernel weights learned by MKLdiv-dc; (f) kernel weights learned by MKLdiv-conv.