以半连续隐含马尔可夫模型HMM 为例,分析了影响Viterbi算法效率的主要原因,讨论了提高该算法效率的3种实用策略,即码本剪枝、Beam剪枝及降低精度,在保证一定识别率的前提下,使得搜索处理更加有效,提高了搜索的速度-to semi-continuous HMM HMM example, Analysis of the impact on the efficiency of the Viterbi Algorithm main discussed increasing the efficiency of the algorithm for three kinds of practical strategy, code pruning, beam pruning and reducing precision, In recognition rate guarantee, under the premise of making the process of searching more efficient and improve the search speed