Table 10 Wrapper methods comparison of predictive performance, number of genes selected and computation time
From: Artificial Intelligence based wrapper for high dimensional feature selection
Dataset | Performance (µ [95% CI]) | |||||
|---|---|---|---|---|---|---|
Predictive performance (RMSE) | Number of genes selected | Computation time (minutes) | ||||
StW | AIWRAP-L | StW | AIWRAP-L | StW | AIWRAP-L | |
TCGA-BLCA | 0.79[0.31,1.27] | 0.78[0.30,1.26] | 4[0,9] | 1[0,3] | 5.9[3.2,8.6] | 12.2[10.1,14.3] |
TCGA-CESC | 1.00[0.84,1.16] | 0.98[0.84,1.13] | 10[7, 13] | 5[4, 6] | 11[7.7,14.2] | 14.6[9.9,19.3] |
TCGA-ESCA | 1.04[0.87,1.20] | 1.00[0.85,1.15] | 11[5, 17] | 8[2, 14] | 7.2[4.9,9.5] | 27.9[3.6,52.2] |
TCGA-HNSC | 0.99[0.82,1.16] | 0.98[0.81,1.15] | 16[12, 20] | 6[3, 9] | 11.4[8.7,14] | 20.3[9.3,31.2] |
TCGA-KICH | 1.03[0.61,1.46] | 0.82[0.39,1.25] | 11[9, 13] | 6[4, 8] | 50.2[24.7,75.7] | 10.6[7.5,13.7] |
TCGA-KIRP | 0.95[0.66,1.24] | 0.95[0.65,1.24] | 19[18, 20] | 15[11, 19] | 10.4[8.8,12] | 41.1[12.5,69.8] |
TCGA-LUAD | 1.02[0.93,1.11] | 1.02[0.94,1.09] | 25[22, 28] | 21[16, 26] | 11.6[9.1,14.1] | 42.3[11.6,72.9] |
TCGA-LUSC | 0.99[0.91,1.08] | 0.99[0.91,1.08] | 2[1, 3] | 1[0,2] | 5.7[4.4,7] | 12[8.8,15.2] |
TCGA-PAAD | 1.26[0.74,1.79] | 1.24[0.75,1.73] | 22[20, 24] | 14[9, 19] | 10.8[7.6,14.1] | 29[0.6,57.4] |