Applying the Word Acquiring Algorithm to the Pinyin-to-Character Conversion

This paper applies the information entropy based word acquiring algorithm to the task of Pinyin-to-character (PTC) conversion, which adopts artificial immune network model. Firstly, the artificial immune network is used to overcome the sparse data problem and the independent identical distribution (iid.) assumption. Secondly, the word acquiring algorithm based on information entropy is presented to collect the Chinese word and some typically combinations. The experiments show that our method can achieve a better performance than the n-gram language model, and this kind of improvement is hardly acquired by the classical supervised learning models. In addition, the word acquiring method is applied, and further improves the PTC performance.