Table 8.
RMSE performance of different wrapper methods on the real studies for test data
| Methods | Performance (RMSE) | |||
|---|---|---|---|---|
| Marginal Model Scenarios | ||||
| I | II | III | IV | |
| Mean (95% Confidence Interval) | ||||
| ALASSO | 0.95 (0.95–0.96) |
3.76 (3.67–3.84) |
3.08 (3.01–3.14) |
0.86 (0.81–0.90) |
| LASSO |
0.96 (0.95–0.97) |
3.75 (3.65–3.85) |
3.10 (3.03–3.16) |
0.84 (0.8–0.87) |
| SPLS |
0.97 (0.95–0.99) |
3.61 (3.54–3.69) |
3.35 (3.03–3.66) |
0.77 (0.76–0.79) |
| Enet |
0.95 (0.94–0.96) |
3.79 (3.7–3.87) |
3.15 (3.08–3.23) |
0.85 (0.81–0.90) |
| AEnet |
0.96 (0.94–0.97) |
3.76 (3.67–3.85) |
3.11 (3.07–3.15) |
0.84 (0.8–0.87) |
| AIWRAP-L |
0.94 (0.93–0.94) |
3.65 (3.59–3.71) |
3.02 (2.98–3.06) |
0.83 (0.8–0.86) |
| AIWRAP-LLr |
0.96 (0.94–0.97) |
3.59 (3.55–3.64) |
2.97 (2.91–3.03) |
0.75 (0.73–0.78) |
| AIWRAP-LR |
0.95 (0.94–0.96) |
3.80 (3.72–3.87) |
3.19 (3.11–3.28) |
1.20 (1.17–1.24) |
| Methods | Interaction Model Scenarios | |||
| I | II | III | IV | |
| Mean (95% Confidence Interval) | ||||
| ALASSO |
0.94 (0.93–0.95) |
3.69 (3.61–3.76) |
3.12 (3.02–3.23) |
0.52 (0.49–0.55) |
| GLASSO |
1.44 (1.2–1.68) |
4.46 (4.35–4.57) |
8.24 (5.37–11.11) |
0.31 (0.28–0.34) |
| LASSO |
0.95 (0.94–0.96) |
3.74 (3.67–3.81) |
3.15 (3.02–3.27) |
0.43 (0.39–0.47) |
| SPLS |
1.03 (0.91–1.15) |
3.81 (3.76–3.86) |
4.34 (3.26–5.42) |
0.24 (0.22–0.26) |
| Enet |
0.94 (0.93–0.95) |
3.78 (3.72–3.84) |
3.24 (3.13–3.34) |
0.44 (0.4–0.48) |
| AEnet |
0.93 (0.92–0.94) |
3.73 (3.65–3.81) |
3.14 (3.06–3.21) |
0.53 (0.5–0.56) |
| AIWRAP-L |
0.94 (0.92–0.95) |
3.58 (3.53–3.63) |
3.07 (2.98–3.17) |
0.29 (0.26–0.33) |
| AIWRAP-LLr |
1.04 (0.99–1.1) |
3.76 (3.58–3.93) |
3.65 (3.26–4.04) |
0.26 (0.21–0.31) |
| AIWRAP-LR |
0.93 (0.92–0.94) |
3.70 (3.64–3.76) |
3.22 (3.18–3.26) |
1.11 (0.99–1.24) |
Values in Bold means best results