Chinese Word POS Tagging with Markov Logic

Because Chinese language has many unique characteristics, this leads to many conventional machine learning approaches of word POS tagging to be unsatisfactory and inefficient when applying to Chinese POS tagging task. Accordingly, the performance of the state-of-the-art Chinese word POS tagging is not as good as that for other languages such as English. In this paper, we present a novel Chinese word POS tagging method. We first assume that the characterbased Chinese sentences are segmented to words completely before hand and phonetic notation - pinyin of these corresponding words is also accomplished. Then, we employ Markov Logic Networks (MLNs) to identify the Chinese word POS tags. It can describe or represent easily and flexibly these rich Chinese grammar structure. Furthermore, for considering the requirements of experiment and contrast, we build two benchmark datasets i.e., dataset1 and dataset2. Both datasets correspond to two different sentence types. The dataset1 and dataset...

• Type: Journal
• Year: 2015
• Where: PAISI
• Authors: Zhihua Liao, Qixian Zeng, Qiyun Wang
• Keywords: PAISI 2015, Security Privacy