Research Article
[Retracted] Unsupervised Anomaly Detection Based on Deep Autoencoding and Clustering
Table 4
Experimental results on Pen_global dataset.
| | Pen_global | Precision | F1 | Time (s) |
| K-Means | 0.8780 | 0.6261 | 0.0680 | PCA + K-Means | 0.8660 | 0.6914 | 0.5920 | ICA + K-Means | 0.8050 | 0.6860 | 0.1326 | MDS + K-Means | 0.9333 | 0.6335 | 0.8024 | DAE + K-Means | 0.9849 | 0.7716 | 0.7817 |
| DBSCAN | 0.8799 | 0.9315 | 0.0256 | PCA + DBSCAN | 0.8832 | 0.8612 | 1.4397 | ICA + DBSCAN | 0.8917 | 0.9333 | 1.0609 | MDS + DBSCAN | 0.1373 | 0.2240 | 1.7177 | DAE + DBSCAN | 0.9420 | 0.9675 | 0.44910 |
| Mean-Shift | 0.7867 | 0.5339 | 0.0855 | PCA + Mean-Shift | 0.8478 | 0.6555 | 0.5777 | ICA + Mean-Shift | 0.6667 | 0.4000 | 0.2387 | MDS + Mean-Shift | 0.7778 | 0.6471 | 0.9347 | DAE + Mean-Shift | 0.9265 | 0.8097 | 0.0865 |
|
|