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

Algorithm 1

The Blocked Gibbs Sampling Algorithm.

Input:

 Random initial corpus segmentation

Output:

  Unsupervised co-segmentation of the corpus according to the model

() for each   do

()  for each blingual word-pair   do

()   for co-segmentation of of   do

()    Compute probability ;

()    Where is the distribution of all s that have been generated before except that from ;

()   end for

()   Sample a from the distribution of ;

()   Update counts;

()  end for

() end for