Statistical uyhur POS tagging with TAG predictor for unknown words
Shengwei Tian, Turgun Ibrahim, Hasan Umal, Long Yu · 2009
Automatic text tagging is an important component in higher level analysis of text corpora, and its output can be used in many natural language processing applications. Trigrams tags is an efficient statistical part-of-speech tagging. This paper describes a POS tagging for Uyhur text based on hidden Markov model using trigrams tags. We describe the basic model of Trigrams Tags, the techniques used for smoothing to address the sparse data problem and a tag predictor for unknown words. A comparison has shown that our approach performs significantly for the Uyhur tested corpora.