Application Of Hidden Markov Model In Part-of-Speech Tagging

In this paper,we first introduce the Hidden Markov Model,and its applications in Natural Language Processing.Then,we focus on its application on Part-of-Speech Tagging,and we build our own model based on the Penn Treebank corpus,and achieved an accuracy rate of 90.48%.After that,we survey the smoothing techniques of language modeling,compared some of the most popular algorithms: Plus one,Good-Turing,Jelinek-Mercer and Katz algorithm.Finally,we choose the Katz algorithm to add to our model,and improved the accuracy rate to 91.91%.