A method and a machine for generating a dictionary of address phrase expressions.
One embodiment of the invention includes an apparatus for generating a dictionary
of target phrases that includes an input interface for receiving as its input a
first address phrase included in a list of address phrase expressions, a memory
for storing a dictionary of address phrase variants, including rules for generating
variants of address phrase expressions, and a processing device for generating
variants of address phrases which generates a second address phrase which is different
in expression from the first address phrase to output the second address phrase
to a storage device holding the dictionary of target phrases, based on the input
first address phrase, and the knowledge of rules about variants included in the
dictionary of address phrase variants.