Algorithm 1: Feature Selection Algorithm

Input: $S$—Data sample with features $X,\left|X\right|=d$ $J$—Evaluation measure to be maximized $GS$—Successor generation operator Output: Solution Begin: Solution: = (weighted) feature subset                   $L≔start\_point(X);$ Solution: = $\left\{best of L according to J\right\};$ Repeat:                $L≔search\_strategy(L,GS\left(J\right),X);$                $X^{\prime }≔\left\{best of L according to J\right\};$ $if J\left(X^{\prime }\right)\ge J\left(Solution\right) or \left(J\left(X^{\prime }\right)=J\left(Solution\right) and \left|X^{\prime }\right|<\left|Solution\right|\right)then$                    $Solution=X^{\prime };$ Until Stop (J, L).