Skip to main content

Table 5 Average empirical power

From: Unsupervised gene set testing based on random matrix theory

# Cov. struct, n, σ 2, ρ TWT MPDT MLRT SGSE
MVN-1 Single block, 100, 1, 0.1 0.87 0.35 0.79 0.76
MVN-2 Single block, 50, 1, 0.1 0.52 0.20 0.42 0.55
MVN-3 Single block, 20, 1, 0.1 0.24 0.16 0.18 0.37
MVN-4 Single block, 100, 1, 0.15 0.99 0.66 0.97 0.95
MVN-5 Single block, 100, 1.1, 0 0.41 0.38 0.16 0.10
MVN-6 Single block, 100, 1.15, 0 0.63 0.68 0.29 0.10
MVN-7 Multi-block, 100, 1, 0.2/0 0.31 0.28 0.56 0.19
MVN-8 Anti-cor. multi-block, 100, 1, 0.1/-0.1 0.85 0.32 0.74 0.06
MVN-9 Repeated block, 100, 1, 0.1 0.49 0.11 0.23 0.10
Binom-1 Single block,100, 1, 0.1 0.64 0.02 0.26 0.55
  1. Average empirical power for each of the evaluated competitive methods computed on 1000 simulated data sets for the ten simulation designs detailed in Table 3. The largest average power found for each design is listed in bold