AI_Chp_HMM_(byPWKXCF)隐马尔可夫.ppt

第九讲 隐马尔可夫模型原理及应用Chapter 9 Hidden Markov Models (Part B) Weike Pan, and Congfu Xu {panweike, xucongfu}@ Institute of Artificial Intelligence College of Computer Science, Zhejiang University October 12, 2006 References Bonnie Dorr, Christof Monz. Lecture of Hidden Markov Models for CMSC 723: Introduction to Computational Linguistics, Spring 2004. /~bonnie/courses/cmsc723-04/ Xuedong Huang, Alex Acero, Hsiao-Wuen Hon. Spoken Language Processing : A Guide to Theory, Algorithm, and System Development. Prentice Hall PTR, Upper Saddle River, N.J., 2001. The HTK Book (for HTK Version 3.3) http://htk.eng.cam.ac.uk/ HTK is a popular toolkit for building Hidden Markov Models (HMMs). Outlines Preparations Three Basic Problems Applications Insights of HMM Outlines Preparations The framework of Bayesian Theory Markov Assumption Representations of HMM Three Basic Problems Applications Insights of HMM The framework of Bayesian Theory Markov Assumption The Markov assumption states that probability of the occurrence of word wi at time t depends only on occurrence of word wi-1 at time t-1 Chain rule: Markov assumption: Representations of HMM States: A set of states S=s1,…,sn Transition probabilities: A= a1,1,a1,2,…,an,n Each ai,j represents the probability of transitioning from state si to sj. Emission/output probabilities: A set, B, of functions of the form bi(ot) which is the probability of observation ot being emitted by si Initial state distribution: is the probability that si is a start state Outlines Preparations Three Basic Problems The Evaluation Problem: Compute the Probability of an Observation Sequence The Decoding Problem: Obtain Maximum Likelihood State Sequence The Learning Problem: Estimate HMM parameters Applications Insights of HMM Problem 1: The Evaluation Problem Problem 1: Given the model and an observation sequence O=o1,…,oT , what is the probability ? i.e. the probability of the model that generates the ob