A Discernibility Matrix is a matrix in which the classes are indecies. In the matrix, the (condition) attributes which can be used to discern between the classes in the corresponding row and column are inserted.
From the previous example, one can observe that the only
attribute with a different value for classes and is
studies. This attribute is therefore placed in its corresponding
places in the matrix. Naturally, the matrix will be symmetric due to
the fact that the attributes that differs in value for objects a and
b, differs ``the other way around'' in value for b and a. Completing
the calculation of the discernibility matrix results in the matrix
shown in Table 
.
| - | studies | education, works | |
| studies | - | studies, education, works | |
| education, works | studies, education, works | - |
If some of the classes have the same decision value, one might decide not to discern between these classes. By doing so, attributes are not added to the matrix for classes with the same decision value. This can result in more simplistic rules if any classes have the same decision value. In the example IS presented earlier this is not an option, since all classes have different decision values.