Eiichiro Sumita and Hitoshi IIda
This paper proposes Example-Based NLP (EBNLP) which uses EXAMPLES (pairs of types of data, e.g. a sentence and its parse) extracted from a corpus and the DISTANCE between examples, explaining Example-Based Machine Translation (EBMT) in particular. EBMT prototype has been implemented to deal with frequent and polysemous linguistic phenomena such as Japanese verbs, case particles and English prepositions. The average success rate of each phenomenon and interesting relationships between success rate and example database parameters are presented.