Table 3.
Schwarz Bayesian Criterion (SBC) for models of increasing complexity (lower SBC values indicate a better solution), using data from the development set. Adding smoothing parameters for skewness and kurtosis parameters does not improve model fit
| GAMLSS Distribution | Parameters with smoothing splines | |||
|---|---|---|---|---|
| Median | Median, Variance | Median, Variance, Skewness | Median, Variance, Skewness, Kurtosis | |
| NO | 9343.77 | 9309.87 | – | – |
| GA | 9503.95 | 9379.81 | – | – |
| TF | 9333.56 | 9313.83 | 9312.84 | – |
| BCCG | 9334.84 | 9281.30 | 9292.47 | – |
| BCT | 9341.22 | 9288.37 | 9299.56 | 9306.63 |
| BCPE | 9341.87 | 9285.50 | 9295.17 | 9301.96 |
Abbreviations: NO Normal, GA Gamma, TF t-Family, BCCG Box Cox Cole and Green, BCT Box Cox t-distribution, BCPE Box Cox Power Exponential