By creating rules using different sets of the attributes one will get
default rules fitting many objects. Still, this may be confusing. Take
the following example:
| Node: | Rule: |
| (A B) | A=1 B=2 D=4 with probability 0.90 |
| (A ) | A=1 D=3 with probability 0.90 |
| (B) | B=2 D=4 with probability 0.6 |
How are we going to decide which decision value is correct for an object matching all these rules? We believe some kind of voting described in the previous section would be the best way of solving this particular question.