RSS AI Zone
I’ve incorporated the new transform functions into Morti’s script, and while he’s a little slow to respond when he has to make the transforms, he seems to be doing so flawlessly, provided that the words in the input are also found in the list of verbs/adverbs that he uses. Unfamiliar words may still trip him up, but that’s a rarity now.
My next step is to use some sort of ontology (e.g. WordNet, or some other) to determine the part of speech for each word, and only do a transform under the proper circumstances. My biggest problem is that, in looking at the MySQL version of WordNet, I’m coming to understand the “holes” in my experience regarding SQL queries, and extracting data from a complex database architecture. I’ve also briefly considered creating my own “crude ontology”, but I really don’t know what information other than POS and whether the word has a “root” (quickly -> quick) or a “stem/stems” (quick -> quicker/quickest, etc.).
Time to head back to class, I guess. 
