Algorithm 3 Data preparation for training of models $M_i$

Input: $S″=\{S_{i,t}:i=1\cdots n,t=1\cdots m\}$, $A″=\{a_t:t=1\cdots m\}$, M, $i^{*}$  Output: $D=\{d_{i,t}:i=1\cdots n,t=1\cdots m\}$ 1:for$t=1\cdots m$do 2:    $found=$ false 3:    $k=0$ 4:    repeat 5:        $X=$ set of all k-element subsets of $\{\{1,2,\cdots ,n\}\setminus \left\{i^{*}\right\}\}$ 6:        repeat 7:           $Z=$ random element from X 8:           $X=X\setminus Z$ 9:           $Z=Z\cup \left\{i^{*}\right\}$ 10:           $S_t=\{S_{i,t}:i\in Z\}$ 11:           if $R(S_t,M)=a_t$ then 12:               $found=$ true 13:           end if 14:        until $found$ or $X=\varnothing$ 15:        $k=k+1$ 16:    until $found$ or $k=n-1$ 17:    for $i=1\cdots n$ do 18:        if $found$ and $i\notin Z$ then 19:           $d_{i,t}=0$ 20:        else 21:           $d_{i,t}=1$ 22:        end if 23:    end for 24:end for