Algorithm 1: Pseudo-Code of IZOA-LightGBM

Start the IZOA-LightGBM model. Input: Dataset. Perform data preprocessing. Initialize the LightGBM model. Define the IZOA fitness function using Equation (1). Set the number of iterations ($T$), the number of zebras’ population ($z$), and the dimension of the problem to be optimized for the IZOA. Use EOBL, i.e., Equations (6) and (7), to create the initial zebra population. Evaluate the fitness values of the zebras. for $k=1 : T$   Update the Pioneer Zebra $\left(PZ\right)$   for $i=1 : z$    #Phase 1    Calculate the new position of the $i$th zebra using Equation (2).    Update the position of the $i$th zebra using Equation (3).    #Phase 2    $p=rand\left(\right)$    if $p\le 0.5$:     Calculate the new position of the $i$th zebra using $M_1$ of Equation (4).    else:     Calculate the new position of the $i$th zebra using $M_2$ of Equation (4).    end if    Update the position of the $i$th zebra using Equation (5).    Use the FDS, i.e., Equations (8) and (9), to update the position of the $i$th zebra again.    Save the $PZ$.   end for $i=1 : z$ end for $k=1 : T$ Input the $PZ$ obtained by the IZOA into the LightGBM model. Obtain the LightGBM model with optimal hyperparameters. Evaluate the IZOA-LightGBM model. Use the optimal IZOA-LightGBM model for detection. End the IZOA-LightGBM model.