I have a large set of data, and two distinct ways of labelling that data, that aren't inherently linked but I suspect they may be. Each method takes a piece of data and assigns it one or more labels, but there are no labels that overlap between the method. For example:
Item Labels (A) Labels (B)
Cow (4 legs), (milk) (white)
Chair (4 legs) (brown), (sit)
Coin (value), (small) (circle), (metal)
Fridge (milk), (big) (metal), (white)
Obviously this is a silly example but I just want to illustrate what I have.
What I want to do is some test to see if they are dependent, in that for each label from Labels (A) I want to calculate a group of labels from Labels (B) that it is associated with. For example the label (milk) from A is associated with (white) from B, as they "commonly" coincide on the same item.
What's a good way to do this?
