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

Figure 6.

Figure 6

Training time analysis of the HHO-DBN model under the epsilon dataset. The results portrayed that CPIO-FS with the HHO-DBN technique resulted in the least training runtime compared to the other methods. For instance, with no FS, the HHO-DBN technique has gained a lower runtime of 318.54 s, whereas the SVMC, LRC, and NBC techniques have accomplished a higher runtime of 400.38 s, 430.48 s, and 340.42 s, respectively. Moreover, sequential CHC, the HHO-DBN approach has attained a minimal runtime of 265.98 s, whereas the SVMC, LRC, and NBC approaches have accomplished superior runtimes of 345.27 s, 398.07 s, and 300.21 s, respectively. Furthermore, MR-EFS, the HHO-DBN method has reached an increased runtime of 243.09 s, whereas the SVMC, LRC, and NBC methodologies have accomplished a higher runtime of 334.18 s, 367.29 s, and 264.26 s, respectively. Finally, CPIO-FS, the HHO-DBN method has attained a lower runtime of 218.96 s, whereas the SVMC, LRC, and NBC algorithms have accomplished superior runtimes of 298.36 s, 312.78 s, and 220.90 s, respectively.