带驻留时间HMM2的Forward-Backward算法  

Forward-Backward Algorithm for state duration HMM2

作　　者：杜世平[1] 汪建[1] 

机构地区：[1]四川农业大学生命科学与理学院,四川雅安625014

出　　处：《重庆工商大学学报（自然科学版）》2007年第1期9-11,共3页Journal of Chongqing Technology and Business University:Natural Science Edition

基　　金：国家自然科学基金资助项目(30300219)

摘　　要：讨论了用状态驻留时间来模型化传统HMM2模型,对传统HMM2的状态转移和输出观测值的Markov假设条件作了改进,在新模型的转移概率和输出观测值的概率中加入驻留时间,并在传统HMM2的基础上定义了新模型的前向-后向变量算法,导出了新模型的前向-后向算法的迭代公式,以及在给定模型λ的条件下,产生观测序列O的概率计算公式.　This paper discusses that state duration is used to model traditional HMM2 model,improves Markov assumption condition of state transition and output observation of traditional HMM2,adds state duration to transition probability of new model and probability of output observation,defines forward-backward variables algorithm of the new model based on traditional HMM2 and deduces the iterative formula of Forward-Backward Algorithm of the new model and probability calculation formula of observation sequence O under the model λ.

关 键 词：二阶隐马尔可夫模型 前向-后向算法 驻留时间 

分 类 号：O212.7[理学—概率论与数理统计]