The LINGUIST List is dedicated to providing information on language and language analysis, and to providing the discipline of linguistics with the infrastructure necessary to function in the digital world. LINGUIST is a free resource, run by linguistics students and faculty, and supported by your donations. Please support LINGUIST List during the 2017 Fund Drive.
|Title:||An Integrated Approach to Chinese Word Segmentation|
|Email:||click here TO access email|
|Linguistic Field:||Computational Linguistics|
|Abstract:||This paper presents an integrated approach to Chinese word segmentation. In this work, known word segmentation and unknown word identification are taken as one integrated process instead of two separate tasks. Furthermore, both known and unknown word candidates are scored equally with a hybrid model, in which three types of features are statistically computed and incorporated, including contextual information (viz. modified word bigram models), affinity relations between the components of words (viz. inside and outside word-juncture models), and the internal word-formation power of components in a word (viz. the word-formation patterns). To enhance the precision and avoid the problem of combination explosion in the construction of word candidates, in particular the unknown word candidates, a filter algorithm is used to block some ineligible unknown word candidates. In this way, ambiguities and unknown words can be resolved effectively. The results of our experiment on the Peking University corpus show that the proposed approach outperforms some other methods.|
Add a new paper
Return to Academic Papers main page
Return to Directory of Linguists main page