Table 10.
Numerical performance of the proposed 3-Layer DDCAE of Table 2 when it is trained according to the layer-wise approach of Vincent et al. (2010) under the Normal dataset. The same simulation setup of Section 5.1 at is considered under the Normal datasets.
Distance | Accuracy | Precision | Recall | F-measure | AUC |
---|---|---|---|---|---|
Kullback–Leibler | 99.61 | 0.9961 | 0.9960 | 0.9960 | 0.9992 |
Bhattacharyya | 99.22 | 0.9923 | 0.9921 | 0.9922 | 0.9991 |
Euclidean | 92.93 | 0.9401 | 0.9292 | 0.9346 | 0.9580 |