Table 3 Feature selection performance of different approaches in simulated scenarios for marginal models
From: Artificial Intelligence based wrapper for high dimensional feature selection
Methods | Performance (Number of features selected) | |||
|---|---|---|---|---|
Marginal Model Scenarios | ||||
1_M | 2_M | 3_M | 4_M | |
p = 50 | p = 50 | p = 100 | p = 100 | |
Target Features: 10 | Target Features: 10 | Target Features: 10 | Target Features: 10 | |
Mean (Range) | ||||
ALASSO | 24 (18–32) | 16 (11–35) | 27 (20–39) | 28 (14–46) |
LASSO | 25 (18–37) | 23 (14–40) | 32 (16–57) | 33 (14–55) |
SPLS | 23 (14–35) | 16 (10–39) | 25 (12–50) | 19 (11–47) |
Enet | 27 (18–36) | 25 (14–41) | 32 (21–45) | 32 (17–55) |
AEnet | 26 (21–30) | 18 (11–35) | 28 (20–43) | 30 (15–48) |
AIWRAP-L | 29 (24–33) | 24 (19–31) | 44 (29–59) | 44 (34–51) |
AIWRAP-LLr | 15 (11–22) | 16 (10–31) | 18 (10–26) | 19 (10–45) |
AIWRAP-LR | 12 (10–16) | 12 (10–16) | 14 (10–21) | 13 (10–22) |