Algorithm 1 Weight learning.
1: functionWeight Learning($X,Y,\alpha ,NumEpochs$)
2:    $w\leftarrow [1,1,\cdots ,1]$
3:    $minLoss\leftarrow \infty$
4:    $\widehat{w}\leftarrow w$
5:    for $epoch\leftarrow (1,NumEpochs)$ do
6:        $\widehat{Y}\leftarrow \left[\right]$
7:        for$x_i\in X$ do
8:           $S\leftarrow Remove(x_i,X)$	▹ Remove $x_i$ from training set
9:           $\widehat{y_i}\leftarrow EVREG(x_i,S,w)$	▹ Predict example $e_i$ with training set S
10:           $Insert(\widehat{y_i},\widehat{Y})$	▹ Insert $\widehat{y_i}$ in $\widehat{Y}$
11:        end for
12:        $loss\leftarrow L(Y,\widehat{Y})$	▹ Compute loss
13:        $gradient\leftarrow CalculateGradient(Y,\widehat{Y})$	▹ Compute gradient
14:        $w\leftarrow w+\alpha *gradient$	▹ Update weights vector
15:        if $minLoss>loss$then	▹ Save best weight vector
16:           $minLoss\leftarrow loss$
17:           $\widehat{w}\leftarrow w$
18:        end if
19:    end for
20:    return $\widehat{w}$
21: end function