Algorithm 2 Sparse linear explanation using LIME.
Require: Classifier f; Features number K; Instance x to be explained; Similarity kernel ${\pi }_x$
$\mathcal{Z}=\left\{\right\}$
${\mathcal{X}}_N$ = SAMPLE_AROUND$\left(x^{\prime }\right)$
for $x_i$ in ${\mathcal{X}}_N$ do
$\mathcal{Z}=\mathcal{Z}\cup (x_i,f\left(x_i\right),{\pi }_x\left(x_i\right))$
end for
$\omega$ = K-Lasso($\mathcal{Z},\mathcal{K}$)
return $\omega$	▹ Explanation for an individual predict