触发式语言模型下的混淆网络解码方法

Decoding method integrating of confusion network based on Trigger language model

将触发式语言模型应用于混淆网络解码过程来提高汉字识别率。为了利用词间的长距离依赖信息,提出了基于词义类对触发式语言模型的混淆网络解码方法。实验结果显示,该方法可以使汉字错误率相对下降7.9%。

The decoding method integrating of confusion network is studied.Trigger language model based on semantic class pairs is proposed to model dependence relationship between long-span words.The model is integrated with confusion network decoding process.Different speech recognition systems utilize different knowledge sources and modeling methods,consequently their error pattern is also different.Experimental results show the method can relatively reduce character error rate by 7.9% respectively.