【摘要】 拼音到汉字转换一直是中文信息处理研究的一个重要方面。在前人的基础上改进了他们所使用的数据平滑算法,给出了两种应用于不同场合的公式,使得在结果中能够得到不同要求的解。同时,给出了利用隐马尔可夫模型(HMM)来设计实现这样一种转换,并且利用Viterbi算法来对模型进行求解的设计细节与方法。最后利用随意选择的语料库来对计算机进行训练学习并给出了测试结果,同时分析了错误结果的原因。更多还原

【Abstract】 The conversion of Pinyin into Chinese characters is an important aspect of Chinese information processing. This paper gives improvement on cache-based natural language algorithms on the basis of predecessors, with two formulas applied to two different occasions to meet different requirements. Also it shows how to achieve such a conversion by using hidden Markov model (HMM) and Viterbi algorithm is used to solve the model. This article gives a detailed way about the design. At the end we randomly choose the corpus to train on the computer and give test results with the analysis of the wrong reasons.更多还原