packages feed

NaturalLanguageAlphabets-0.1.1.0: scoring/simpleunigram.score

-- !!! THIS FILE ASSUMES UTF-8 ENCODING !!!
--
-- These are all vowels, but all 'a's are equivalent with respect to this
-- scoring. It is only required to give 'EqSet's if you have more than one
-- equivalent character. In this EqSet, everything gets a score of 'Eq Vowel'
-- (==2).
EqSet Vowel a â ã
EqSet Vowel e
EqSet Vowel i
EqSet Vowel o
EqSet Vowel u
-- These are sets of characters belonging to a certain class. In this case, we
-- have consonants.
Set Conso    b c d f g h j k m n p q s t v w x y z
Set Liquid   l r
Set Vowel    a\' e\' i\' o\' u\'   ə ɐ æ œ   æh ɔh ɒ ɔ
-- Here we say that whenever we see a consonant aligned to the same consonant
-- in the other word, give this score.
Eq Conso   4
Eq Liquid  3
Eq Vowel   2
-- Whenever we align two consonants that are not equal (or equivalent) we give
-- this score.
InSet   Conso    Conso     0
InSet   Conso    Liquid   -1
InSet   Conso    Vowel    -999999
InSet   Liquid   Conso    -1
InSet   Liquid   Liquid   -1
InSet   Liquid   Vowel    -1
InSet   Vowel    Conso    -999999
InSet   Vowel    Liquid   -1
InSet   Vowel    Vowel     0
-- How to score a gap in a sequence
Gap         -4
-- Later on we want affine gap scoring, it's already here but not used
GapOpen     -4
GapExtend   -4
-- Some types of grammars have differently scored prefixes and suffixes
PreSufOpen   -1
PreSufExtend  0
-- This score is used when we align the same character but it's one we didn't
-- specify in our sets
Match        1
-- And finally how to score a mismatch that doesn't fit into the above
Mismatch    -999999