Hyperparameter optimization method based on dynamic Bayesian with sliding balance mechanism in neural network for cloud computing

Journal of Cloud Computing

Advances, Systems and Applications

Table 3 Comparation of search efficiency

Optimization Method	Results in different search progress (AP50 ± Variance)
Optimization Method	1/4	2/4	3/4	1
GP (baseline)	0.454 ± 0.010	0.459 ± 0.008	0.463 ± 0.004	0.468 ± 0.005
TPE	0.453 ± 0.002	0.460 ± 0.006	0.463 ± 0.005	0.469 ± 0.004
BOHB	0.459 ± 0.003	0.460 ± 0.004	0.464 ± 0.005	0.470 ± 0.002
Dynamic + Hau-PI	0.453 ± 0.003	0.461 ± 0.005	0.465 ± 0.002	0.477 ± 0.002