Research Article

Dictionary Learning Based on Nonnegative Matrix Factorization Using Parallel Coordinate Descent

Table 2

Comparisons of ā€‰( )-based sparsity, Hoyer's sparsity (based on (13)), relative error (based on (14)), and runtime for SNMF, NMFSC, NMF -H, and PCDDL.

Algorithm ā€‰( ) Sparsity Relative error Time (s)

SNMF 96.65 0.4314 0.9904 940
NMFSC 8.00 0.9490 0.2520 415
NMF -H 8.33 0.8957 0.1852 662
PCDDL 8.00 0.9447 0.2925 28