jjohhn has asked for the wisdom of the Perl Monks concerning the following question:
The software detects equivalence ("red car has_brand chevy" <==>"car has_color red, has_brand chevy") The notation for this usually uses logic symbols, but I don't want to clutter this post right now.
The software does something else too (calculates subsumed children), but understanding how to do just the first piece would satisfy me for now.
The terminology would require a few disjoint hierarchies; one for the relationship types (is_a, has_color, has_brand), one for the objects (vehicle perhaps), and one or for the values of the attributes (red, blue, chevy, ford) The attributes in this example would be in two hierarchies: Colors and Brands.
My big questiojn is how do I make a data structure to do this? I am nearly sure it si not done though tradition OO clases, as I don't think that could scale to the level of the commercial product (hundreds of thousands of concepts, over a million relationships talking into a big oracle database.
I just can't get my mind around the first steps of how to model this on a tiny scale in perl. I'd like to output to a set of tables because I am most familiar with that; the commercial product outputs XML that is then converted to tables though.
|
|---|
| Replies are listed 'Best First'. | |
|---|---|
|
Re: toy subsumption classifier
by eoin (Monk) on Mar 11, 2003 at 23:34 UTC | |
by jjohhn (Scribe) on Mar 12, 2003 at 01:41 UTC | |
|
Re: toy subsumption classifier
by demerphq (Chancellor) on Mar 11, 2003 at 22:54 UTC | |
by mdupont (Scribe) on Mar 12, 2003 at 16:16 UTC |