Unsupervised learning of part-of-speech tagging using decision trees in English TTS
英文文语转换系统中基于决策树的词性标注的非监督学习

Part-of-Speech is essential for prosody generation module and homograph disambiguation module in English TTS(text to speech) system. An unsupervised learning of part-of-speech tagging using decision trees was described, under the condition of a lexicon. In addition, the problem of unknown word guessing was also discussed.