شماره ركورد كنفرانس :
3502
عنوان مقاله :
IBM Word-Alignment Model 1 for Statistical Machine Translation
Author/Authors :
N Daneshgar Ferdowsi University of Mashhad , M Sarmad Ferdowsi University of Mashhad
كليدواژه :
word alignment , statistical models , IBM model 1 , EM algorithm
عنوان كنفرانس :
چهل و پنجمين كنفرانس رياضي ايران
چكيده لاتين :
This paper discribes the word alignment based on IBM model 1 for an English-Persian Parallel
Corpus (Mizan). We only consider 100,000 sentence pairs. It is been obtained that for each sentence, more
than 80 percentage of word alignments between English and Persian words are reasonable.
