Table 4 Mean log-posterior predictive likelihood for each model under each data-generated condition. Row names correspond to the pattern of covariate inclusion. Column names correspond to the models. Bolded values indicate the best performing model based on a pairwise t-test. If multiple values are bolded, then model performances were not significantly different at 0.01 level
From: A hierarchical spike-and-slab model for pan-cancer survival using pan-omic data
| Â | Hierarchical | Fixed (0.5) | Full model | Shared | Null model | Joint | Separate | Hsaft |
|---|---|---|---|---|---|---|---|---|
All In (Prob = 0.5) | – 7687.97 | – 7771.80 | – 7883.81 | – 7783.47 | – 11662.95 | – 10675.67 | – 7819.95 | – 7893.56 |
All In (Prob = 0.1) | – 8221.97 | – 8310.11 | – 8520.15 | – 8229.85 | – 9729.56 | – 9869.96 | – 8336.16 | – 8359.26 |
Indep. In (Prob = 0.5) | – 7755.32 | – 7751.48 | – 7951.62 | – 7749.98 | – 11701.51 | – 11287.71 | – 7815.04 | – 7908.40 |
Indep. In (Prob = 0.1) | – 7360.53 | – 7421.77 | – 7723.09 | – 7349.98 | – 9580.55 | – 10358.29 | – 7447.81 | – 7488.49 |
All In | – 307.15 | – 8359.16 | – 8291.82 | – 8292.16 | – 12863.43 | – 11096.83 | – 8436.40 | – 8409.30 |
None In | – 7827.13 | – 7918.48 | – 8173.49 | – 7804.41 | – 7800.81 | – 9504.38 | – 7935.25 | – 7921.88 |