Paper: Chinese-Uyghur Sentence Alignment: An Approach Based on Anchor Sentences

ACL ID W09-3108
Title Chinese-Uyghur Sentence Alignment: An Approach Based on Anchor Sentences
Venue Building and Using Comparable Corpora
Session
Year 2009
Authors
  • Samat Mamitimin (Xinjiang University, Urumqi China; Communication University of China, Beijing China)
  • Min Hou (Communication University of China, Beijing China)

This paper, which builds on previous studies on sentence alignment, introduces a sentence alignment method in which some sentences are used as “anchors” and a two step proce- dure is applied. In the first step, some lexical information such as proper names, technical terms, numbers and punctuation marks, loca- tion information and length information are used to generate anchor sentences that satisfy some conditions. In the second step, texts are divided into several segments by using the anchor sentences as boundaries, and then the sentences in each segment are aligned by us- ing a length-based approach. By applying this segmentation technique, the method avoids complex computation and error spreading. Experimental results show that the precision of the method is 94.6% o...