Exact Computation of the Observed Information Matrix for Hidden Markov Models

This article describes a new algorithm for exact computation of the observed information matrix in hidden Markov models that may be performed in a single pass through the data. The score vector and log-likelihood are computed in the same pass. The new algorithm is derived from the forward–backward algorithm traditionally used to evaluate the likelihood in hidden Markov models. Our result is discussed in the context of previous approaches that have been used to obtain approximate standard errors of parameter estimates in these models. Implications for parameter estimation are also discussed. An application of the proposed methods to rainfall occurrence data is provided.

[1]  Kenneth Lange,et al.  Numerical analysis for statisticians , 1999 .

[2]  R J Cook,et al.  Estimation of Operating Characteristics for Dependent Diagnostic Tests Based on Latent Markov Models , 2000, Biometrics.

[3]  Xiao-Li Meng,et al.  Using EM to Obtain Asymptotic Variance-Covariance Matrices: The SEM Algorithm , 1991 .

[4]  L. Baum,et al.  A Maximization Technique Occurring in the Statistical Analysis of Probabilistic Functions of Markov Chains , 1970 .

[5]  Biing-Hwang Juang,et al.  Hidden Markov Models for Speech Recognition , 1991 .

[6]  D. Rubin,et al.  Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper , 1977 .

[7]  James P. Hughes,et al.  Incorporating Spatial Dependence and Atmospheric Data in a Model of Precipitation , 1994 .

[8]  P. Guttorp.  Stochastic modeling of scientific data , 1995 .

[9]  L. Baum,et al.  An inequality and associated maximization technique in statistical estimation of probabilistic functions of a Markov process , 1972 .

[10]  Murray A. Cameron,et al.  Hidden Markov chains in generalized linear models , 1998 .

[11]  F Sachs,et al.  A direct optimization approach to hidden Markov modeling for single channel kinetics. , 2000, Biophysical journal.

[12]  James P. Hughes,et al.  Computing the observed information in the hidden Markov model using the EM algorithm , 1997 .

[13]  Glen A. Satten,et al.  Markov Chains with Measurement Error: Estimating the ‘True’ Course of a Marker of the Progression of Human Immunodeficiency Virus Disease , 1996 .

[14]  K Lange,et al.  Computational advances in maximum likelihood methods for molecular phylogeny. , 1998, Genome research.

[15]  L. Baum,et al.  An inequality with applications to statistical estimation for probabilistic functions of Markov processes and to a model for ecology , 1967 .

[16]  T. Louis.  Finding the Observed Information Matrix When Using the EM Algorithm , 1982 .

[17]  L. Baum,et al.  Statistical Inference for Probabilistic Functions of Finite State Markov Chains , 1966 .

[18]  L. R. Rabiner,et al.  An introduction to the application of the theory of probabilistic functions of a Markov process to automatic speech recognition , 1983, The Bell System Technical Journal.

[19]  P. Guttorp,et al.  A non‐homogeneous hidden Markov model for precipitation occurrence , 1999 .

[20]  Lain L. MacDonald,et al.  Hidden Markov and Other Models for Discrete- valued Time Series , 1997 .

[21]  J. Rice,et al.  Maximum likelihood estimation and identification directly from single-channel recordings , 1992, Proceedings of the Royal Society of London. Series B: Biological Sciences.

[22]  Xiao-Li Meng,et al.  The Art of Data Augmentation , 2001 .

[23]  Pierre A. Devijver,et al.  Baum's forward-backward algorithm revisited , 1985, Pattern Recognit. Lett..