Table 1.
A summary of the model averaged coefficients for the two models, the first consisting of six features with an inclusion fraction >0.8 and the second consisting of eight features with an inclusion probability (relative feature importance) >0.9
| Inclusion fraction | Inclusion probability | Mean coefficient | Weighted mean coefficient | Model | ||
|---|---|---|---|---|---|---|
| (Intercept) | – | – | 0.792 | 1.027 | 1, 2 | |
| APOA4 | IDQNVEELK | 1.00 | 1.000 | −0.238 | −0.320 | 1, 2 |
| APOC3 | GWVTDGFSSLK | 1.00 | 1.000 | −0.287 | −0.334 | 1, 2 |
| HPT | VTSIQDWVQK | 1.00 | 0.998 | 0.266 | 0.287 | 1, 2 |
| IC1 | TNLESILSYPK | 0.93 | 0.996 | 0.141 | 0.196 | 1, 2 |
| APOA2 | SPELQAEAK | 0.89 | 1.000 | −0.248 | −0.390 | 1, 2 |
| ITIH4 | GPDVLTATVSGK | 0.83 | 0.992 | 0.0932 | 0.153 | 1, 2 |
| ANT3 | LPGIVAEGR | 0.67 | 0.965 | 0.0382 | 0.0775 | 2 |
| APOH | EHSSLAFWK | 0.59 | 0.949 | 0.0320 | 0.0622 | 2 |
The mean coefficient is the mean of the coefficients for the feature of interest based on all of the models. The weights used for the weighted mean coefficient are the model probabilities (Akaike weights)