Morphology and Finite-State Transducers
-
What is morphology?
- The internal structure of words: words are composed
of one or more morphemes
- The accompanying semantics: building up the meanings of words
from the meaings of morphemes
-
Why is morphological analysis useful in CL?
- To indicate when two different words have the same root; useful
in information retrieval and other applications for stemming
and term expansion
- To analyze the semantics of morphologically complex words, in any
application that requires sophisticated natural language understanding
or generation
-
Finite-State Transducers (FSTs)
- An FST is a finite-state machine with two "tapes": one for input
and one for output
- Why use FSTs for morphological analysis?
- An example
(2)
back to regular expressions and finite-state automata
(return to syllabus)