Algorithm 3. Bayesian hyperparameter optimization algorithm for the classification algorithm sub-model of the proposed method
Input: Objective function $f\left(x\right)$, search space $\chi$ of hyperparameter, sampling budget n, initial point of hyperparameter $x_0$, probability model P, acquisition function $a\left(x\right)$
Output: Optimal hyperparameter set $x*$ (multi-dimensional vector)
Define: (1) Objective function $f\left(x\right)$ is set to exhibit metric accuracy, as defined by Equation (13) in Section 3.4; (2) Probability model P is set to be Gaussian process model; (3) Acquisition function $a\left(x\right)$ is set to be expected improvement (EI) function.
1	Initialization: $t=0,X_t=\left\{x_0\right\},Y_t=\left\{f\right(x_0\left)\right\}$
2	while ($t<n$) do
3	Obtain ($X_{\mathrm{t}}, Y_t$) by P
4	Calculate the next sampling point $x_{t+1}={\mathrm{a}\mathrm{r}\mathrm{g}max}_{x\in \mathcal{X}}\hspace{1em}a\left(x\right)$
5	Evaluate the objective function $y_{t+1}=f\left(x_{t+1}\right)$
6	Update: $X_{t+1}=X_t\cup \left\{x_{t+1}\right\}$
7	Update: $X_{t+1}=X_t\cup \left\{x_{t+1}\right\}$
8	Update iteration step $t=t+1$
9	end while
10	Output $x*={\mathrm{a}\mathrm{r}\mathrm{g}max}_{x\in \mathcal{X}}\hspace{1em}f\left(x\right)$ as the optimal hyperparameter set $x*$