Table 3.
Average misclassification error rate and area under the receiver operating characteristic curve (AUC) for three class simulated data with 5% missing values and different rates of outliers.
| Methods | 3% Outliers MER (AUC) | 5% Outliers MER (AUC) | 7% Outliers MER (AUC) | 10% Outliers MER (AUC) |
|---|---|---|---|---|
| RF | 5.60 (0.9661) | 10.60 (0.9122) | 12.97 (0.8631) | 21.63 (0.7687) |
| PPCA | 4.33 (0.9718) | 7.73 (0.9375) | 11.03 (0.8797) | 21.47 (0.7734) |
| kNN | 3.67 (0.9783) | 9.40 (0.9251) | 13.80 (0.8634) | 18.60 (0.7785) |
| BPCA | 4.00 (0.9735) | 10.17 (0.9131) | 13.37 (0.8688) | 21.77 (0.7562) |
| EM-PCA | 5.20 (0.9645) | 8.93 (0.9302) | 11.07 (0.8799) | 21.67 (0.7674) |
| Zero | 4.67 (0.9696) | 8.67 (0.9172) | 15.80 (0.8546) | 15.20 (0.8054) |
| Mean | 4.13 (0.9724) | 10.97 (0.8995) | 13.93 (0.8620) | 18.23 (0.7924) |
| Median | 4.20 (0.9721) | 10.03 (0.9152) | 13.50 (0.8673) | 17.93 (0.7948) |
| Minimum | 5.23 (0.9626) | 8.30 (0.9193) | 12.13 (0.8913) | 14.33 (0.8214) |
| rmiMAE | 1.82 (0.9816) | 2.21 (0.9783) | 2.57 (0.9721) | 3.29 (0.9672) |
| Proposed | 1.28 (0.9892) | 1.32 (0.9877) | 1.39 (0.9862) | 1.48 (0.9835) |
Bold indicates the lower MER and Higher AUC throughout the column.