Research Article

A Dirichlet Process Mixture Based Name Origin Clustering and Alignment Model for Transliteration

Table 2

The average length of English character sequence aligned to each Chinese sequence.

ā€‰ Model EO ECJ-O Multi-O

#(Targets) GIZA++ 14.43 5.35 6.62
BBAM 6.06 2.45 2.91
cDPMM 9.32 3.45 4.28