Skip to main content
. 2021 Dec 17;11:24138. doi: 10.1038/s41598-021-03019-y

Figure 5.

Figure 5

AUC analysis of HHO-DBN model under epsilon dataset. The results ensured that the CPIO-HS with the HHO-DBN model offered the maximum outcome over the other techniques. For instance, with no FS technique, the HHO-DBN model attained a higher AUC of 72.70%, whereas the SVMC, LRC, and NBC techniques attained a lower AUC of 59%, 62%, and 64%, respectively. Following the sequential CHC approach, the HHO-DBN manner reached an increased AUC of 74.80%, whereas the SVMC, LRC, and NBC methods obtained minimal AUCs of 65%, 67%, and 68%, respectively. In line with MR-EFS, the HHO-DBN method reached a superior AUC of 78.60%, whereas the SVMC, LRC, and NBC methodologies achieved a minimal AUC of 68%, 7%, and 72%, respectively. Last, with the CPIO-FS technique, the HHO-DBN methodologies reached a superior AUC of 93.90%, whereas the SVMC, LRC, and NBC approaches attained minimal AUCs of 80.10%, 82.80%, and 86.50%, respectively.