CAJ | 학술논문

为了更准确地估计状态聚类前有调三音子的模型参数, 从而提高聚类后捆绑状态的精度及系统的识别性能, 针对汉语连续语音识别中, 有些有调三音子的训练样本数非常少, 而其对应的无调三音子的训练样本数相对较多的情况, 提出用其对应的无调三音子的模型参数进行初始化, 并用最大后验概率准则训练模型。汉语大词汇量连续语音识别实验表明, 该方法可以提高训练语料中稀疏三音子聚类前的模型精度, 从而提高系统的识别性能。
위료경준학지고계상태취류전유조삼음자적모형삼수, 종이제고취류후곤방상태적정도급계통적식별성능, 침대한어련속어음식별중, 유사유조삼음자적훈련양본수비상소, 이기대응적무조삼음자적훈련양본수상대교다적정황, 제출용기대응적무조삼음자적모형삼수진행초시화, 병용최대후험개솔준칙훈련모형。한어대사회량련속어음식별실험표명, 해방법가이제고훈련어료중희소삼음자취류전적모형정도, 종이제고계통적식별성능。
In order to estimate toned triphone's model parameters accurately before state clustering and increase recognition rate, this paper used the corresponding toneless triphone model parameters for the initialization of toned triphones, and applied the maximum a posteriori criterion for model estimation. In the experiment of mandarin large vocabulary continuous speech recognition, </sup> this method can improve the accuracy of sparse triphone's model in the training corpus before clustering and achieve certain increase of recognition rate.