Table 2.
Top ranking models and performance.
| Model | Subset | N | Np | Parameters | Acc | k | Sens | Spec | TP | FN | FP | TN | AUC |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Imp_B | Basic | 211 | 3 | LDL-C + ApoB/ApoA-I + TG/ApoB | 0.84 | 0.67 | 0.91 | 0.86 | 32 | 3 | 7 | 42 | 0.92 |
| RFEct_BL | Basic & Lipoprint | 95 | 4 | TG/ApoB + TC/HDL-C + TC + LDL1 | 0.84 | 0.64 | 0.83 | 0.92 | 10 | 2 | 2 | 23 | 0.91 |
| Sel3 | All | 78 | 5 | LDL1 + ApoC-III + TC/HDL-C + BMI + Age | 0.77 | 0.49 | 0.82 | 0.90 | 9 | 2 | 2 | 18 | 0.89 |
| RFEct_A | All | 78 | 5 | LDL1 + TC + ApoA-II + MIDC + TC/HDL-C | 0.77 | 0.46 | 0.82 | 0.80 | 9 | 2 | 4 | 16 | 0.88 |
| RFE78ct_BL | Basic & Lipoprint | 78 | 5 | TC + TC/HDL-C + MIDB + MIDC + LDL1 | 0.74 | 0.41 | 0.82 | 0.85 | 9 | 2 | 3 | 17 | 0.88 |
| RFE78t_B | Basic | 78 | 2 | LDL-C + ApoB/ApoA-I | 0.81 | 0.59 | 0.82 | 0.85 | 9 | 2 | 3 | 17 | 0.87 |
| Sel1 | All | 78 | 3 | LDL1 + ApoC-III + TC/HDL-C | 0.77 | 0.47 | 0.82 | 0.90 | 9 | 2 | 2 | 18 | 0.87 |
| Imp_AdL | Advance & Lipoprint | 78 | 3 | ApoA-II + ApoC-III + LDL1 | 0.77 | 0.47 | 0.73 | 0.75 | 8 | 3 | 5 | 15 | 0.76 |
| RFE78t_Ad | Advanced | 78 | 5 | ApoA-II + ApoC-II + ApoC-III + sdLDL.Day + BMI | 0.77 | 0.49 | 0.91 | 0.60 | 10 | 1 | 8 | 12 | 0.75 |
| RFE78ct_Ad | Advanced | 78 | 5 | Age + ApoA-II + ApoC-II + ApoC-III + sdLDL.Day | 0.85 | 0.66 | 0.73 | 0.65 | 8 | 3 | 7 | 13 | 0.75 |
N number of individuals, Np number of parameters, Acc accuracy, k Cohen’s kappa coefficient, Sens sensitivity, Spec specificity, TP number of true positives, FN number of false negatives, FP number of false positives, TN number of true negatives, AUC area under the ROC curve.