This repository was archived by the owner on Jul 21, 2024. It is now read-only.
TTS does not always match AIML file internal spelling #29
Description
If I paste "WHAT IS YOUR FAVORITE ABBA SONG" (from music.aiml) then the response is correct.
I think it is a tie between "Waterloo" and "Dancing Queen."
But if I speak the same question, STT gets:
" what is your favourite abba song"
which doesn't match successfully inside of the AIML engine (surely because of US vs UK spelling of "favorite"). I am unsure of the canonical way to handle. Does AIML have provisions for locality spelling differences? if not, do we normalize the corpus (which is mostly US EN) and then maybe introduce a normalizing element into the python code of the skill?