Skip to main content

Table 4 Distance correlations with the best choice of rank r for different matrix factorization methods on nine datasets

From: Sparse data embedding and prediction by tropical matrix factorization

Dataset Rank r STMF NMF PMF
Min. Median Max.
AML 3 0.650 0.831* 0.845 0.636 0.623
COLON 3 0.585 0.647 0.688 0.586 0.707*
GBM 3 0.684 0.702* 0.762 0.325 0.330
LIHC 2 0.493 0.515 0.588 0.311 0.649*
LUSC 3 0.498 0.562 0.731 0.697 0.799*
OV 4 0.420 0.569* 0.601 0.347 0.563
SKCM 3 0.480 0.521 0.605 0.633 0.808*
SARC 3 0.493 0.584 0.610 0.649* 0.588
BIC 3 0.350 0.392 0.531 0.227 0.427*
  1. Result of best method in the comparison between STMF and NMF shown in bold. Best result among all three methods (STMF, NMF, PMF) indicated by asterisk