hiding objects and attributes

[rjoberon]

For certain use-cases it would be helpful if one could hide certain attributes or objects in the context and use the remaining context for analysis.

One particular use-case is this: A user could select an association rule with less then 100% confidence and restrict the objects of the context such that the confidence is 100% (e.g., the rules becomes an implication). This could be done by hiding all objects that have the attributes from the premise but not the ones from the conclusion. Then, a recomputation of rules and the lattice would be necessary, of course.

related to issue #135