Skip-N Mongolian Statistical Language Model
Guo‐Qiang Zhang · Journal of Inner Mongolia University · 2008
Etymas and suffixes are the typical characters of Mongolian.A Mongolian word could usually be divided into an etyma and several suffixes.It is difficult to describe the dependent relationship of long distance between an etyma and suffixes if using the N-gram language model.Based on long-distance dependence,a new kind of language model called Skip-N model is presented,and the relation of N-word separated bigram dependence is given.The technique is realized,to applied to a Chinese-Mongolian machine translation system.The experiments prove that the Skip-N language model can improve the translation result.