Algorithm 1. The pseudo code of the constructed classifier.

Input: Training samples set $Data=\left(x_1,x_2,\dots x_N\right)$ and number of learning rounds T. 1.$D_1\left(x_i\right)=1/m$ % Initialize the weight distribution for all samples 2.For $t=1,2,\dots$T   $h_t\left(x_i\right)=LCE\left(Data,D_t\left(x_i\right)\right)$ % Train a learner $h_t\left(x_i\right)$ from $Data$ using $D_t\left(x_i\right)$ based on the LCE method 3.  ${\epsilon }_t={\sum }_{i=1}^N{D}_t\left(x_i\right)\left[y_i\ne h_t\left(x_i\right)\right]$ % Calculate the error of $h_t\left(x_i\right)$ 4.  If ${\epsilon }_t$ > 0.5 then break 5.  ${\alpha }_t=\frac{1}{2}log\left[\frac{\left(1-{\epsilon }_t\right)}{{\epsilon }_t}\right]$ % Determine the weight of $h_t\left(x_i\right)$ 6.  $D_{t+1}=\frac{D_t\left(x_i\right)}{Z_t}\times \{\begin{array}{cc}{e}^{-{\alpha }_t}& if{y}_i=h_t\left(x_i\right)\\ e^{{\alpha }_t}& otherwise\end{array}$ % Update the weight distribution for all samples 7.End

Output: The proposed classifier $H\left(x_i\right)=sign({\sum }_{t=1}^T{\alpha }_t{h}_t\left(x_i\right))$