Morphological Analysis

• Dictionary Enhancement for ChaSen
  • Add lexical entries, especially names of persons & places
  • Differentiate entries if pronunciation depends on POS tags
    (Ex.) ``は'' $\rightarrow$ [ha], [wa]
• Pronunciation Post-Processor: ChaWan
  • Numericals: ``0'' $\rightarrow$ [zero] [rei],   ``1'' $\rightarrow$ [ichi] [iq]
  • Counters:    ``1本'' [iq poN],      ``2本'' [ni hoN],      ``3本'' [saN boN]
• Lexical entries are differentiated by POS tags ('97 version)
  and also by pronunciations if disambiguated in morph. analysis ('98 version)
  (Ex.) 本+ホン+counter        [h o N]
          本+ポン+counter        [p o N]
          本+ホン+noun(book)   [h o N]