Algorithm 3 Heuristic (M, $RM$ or $log$) for induction of decision rules.

Input: Decision table T with condition attributes and row r

Output: Decision rule $rul$ for T and given row r

$Q\leftarrow \varnothing$;

$T^0(attr,v_{attr})\leftarrow T$;

while $T^j(attr,v_{attr})$ is not degenerate do

select attribute $att{r}_i$ as follows:     • heuristic M selects $att{r}_i$ which minimizes the value $M(att{r}_i,r,d_k)$;     • heuristic $RM$ selects $att{r}_i$ which minimizes the value $RM(att{r}_i,r,d_k)$;     • heuristic $log$ selects $att{r}_i$ which maximizes the value $\frac{\beta (att{r}_i,r,d_k)}{{log}_2(\alpha (att{r}_i,r,d_k)+2)}$;

$Q\leftarrow Q\cup \left\{attr\right\}$;

$T^{(j+1)}\leftarrow T^j(attr,v_{attr})$;

$j=j+1$;

end while

$rul\leftarrow {\forall }_{attr\in Q}(att{r}_i=v_{att{r}_i})\to d_k$;