f(c(1))≥f(c(2)) always holds if c(2) contains c(1).
These three properties can be understood as follows. The first property says that an uncommon cell provides more information (more distinctive). The second property means that the distinctiveness level of a cell is reinforced when the uncertainty level is low. The third property implies that we are more confident to associate incidents when they take the same value on “more” attributes.
Reasons that these properties are “good” desired properties are discussed in details in Ref. [27]. Here we only give some examples for illustration. Consider the following three scenarios.