Nonparametric Model for Inupiaq Morphology Tokenization
We present how to use English translation for unsupervised word segmentation of low resource languages. The inference uses a dynamic programming algorithm for efficient blocked Gibbs sampling. We apply the model to Inupiaq morphology analysis and get better results than monolingual model as well as...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Text |
Language: | English |
Subjects: | |
Online Access: | http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.368.1099 http://aclweb.org/anthology/C/C12/C12-3042.pdf |
Summary: | We present how to use English translation for unsupervised word segmentation of low resource languages. The inference uses a dynamic programming algorithm for efficient blocked Gibbs sampling. We apply the model to Inupiaq morphology analysis and get better results than monolingual model as well as Morfessor output. |
---|