论文部分内容阅读
本文详细介绍了隐 Markov 模型(HMM)的基本概念和计算概率的前、后向算法,并采用 HMM 作了多讲话者(三人)小词汇量汉语单字识别实验。正确识别率为98%。
This paper introduces the basic concepts of Hidden Markov Model (HMM) and the forward and backward algorithms of calculating probability, and uses HMM as a multi-speaker (three) small vocabulary Chinese word recognition experiment. The correct recognition rate of 98%.