Skip to main content
. 2018 May 2;34(19):3365–3376. doi: 10.1093/bioinformatics/bty357

Fig. 1.

Fig. 1.

VT and SIDES, whilst searching for predictive signals, mistakenly give high rank to variables that are purely prognostic, with no predictive signal whatsoever (black bars); whereas, INFO+ correctly assigns them a rank no better than random. This is the average ranking score over 200 simulated datasets generated by model M-1, in the absence of any predictive information θ = 0, sample size 2000 and dimensionality p = 30 biomarkers. The dashed line is the average expected score, representing a ranking by random chance. Anything above can be considered as significant