利用關聯式規則解決台語文轉音糸統中一詞多音之歧異

Applying Association Rules in Solving the Polysemy Problem in a Chinese to Taiwanese TTS System

本論文提出一個利用關聯式規則解決台語文轉音系統中的一詞多音問題，台語中的一詞多音指的是一個詞在不同的場合會有不同的發音，如果不做適當處理，會造成合成之台語語音發音的錯誤，導致合成語音之瞭解度下降。有別於既有之解決一詞多音方法，關聯式規則具有推衍出較具物理意義之規則，可以模擬人們在語言使用的習慣，對於解決一詞多音有相當的助益，本文除了利用關聯式規則外，也使用了E-Hownet來解決訓練語料稀疏的問題。我們針對六種常見的多音詞進行實驗，分別為『上』『下』『不』『你』、『我』、『他』等詞。分別得到正確率為93.99%、94.44%、77.82%、95.11%、96.35%、95.99%。實驗結果顯示，相較於現有的實驗方法，本論文提出的方法對於上述六個多音詞皆達到更高的正確率。

This paper proposed an approach that apply association rules to solve the polysemy problem in a Chinese to Taiwanese TTS system. In Taiwanese, one word possibly has several pronunciations. It leads to that Taiwanese TTS System can't work well if wrong pronunciation is synthesized. Thus, we need to decide which pronunciation is proper under some condition in order to enhance the performance of a Taiwanese TTS System. The approach of association rules has the advantage of similate the usage of language by human beings. We also use the information in E-Hownet to overcome the problem of data sparsity. The experiments focus on six common used words, they are “上”(up), “下”(down), “不”(no), “你”, “我”, and “他”(he). Experiment results show that the proposed approach can achieve higher accuracies than the existing methods.