s-news
[Top] [All Lists]

Re: Hidden Markov Models

To: "'Juan Carlos Ruilova Teran'" <JRTeran@itaubba.com.br>
Subject: Re: Hidden Markov Models
From: Martin Maechler <maechler@stat.math.ethz.ch>
Date: Thu, 4 May 2006 09:50:30 +0200
Cc: <s-news@wubios.wustl.edu>
In-reply-to: <200605031500.k43F0krT028697@mailhost1.u.washington.edu>
References: <125A7D9CB9E1934FB1F6EA7F358E9BF2021BC6A0@BBAMSGP2.corp.bba.com.br> <200605031500.k43F0krT028697@mailhost1.u.washington.edu>
Reply-to: Martin Maechler <maechler@stat.math.ethz.ch>
>>>>> "Eric" == Eric Zivot <ezivot@u.washington.edu>
>>>>>     on Wed, 3 May 2006 08:00:47 -0700 writes:

    Eric> You can estimate hidden markov models using the state
    Eric> space functions in S+FinMetrics 2.0. The state space
    Eric> functions allow Markov regime switching in the state
    Eric> space system matrices. Estimation, filtering and
    Eric> smoothing follows Kim's algorithm (see Kim and Nelson
    Eric> State Space Models with Regime Switching, MIT Press).
    Eric> ez

Alternatively, you upgrade to R,
where I quickly find three different R packages on CRAN (the
Comprehensive R Archive Network) that all have functions for
fitting HMMs :

        Package         Function
        -------         --------
        hmm.discnp      shmm(), sim.hmm()
        klaR            hmm.sop()
        msm             msm()   

the latter stands for   
    'Multi-state Markov and hidden Markov models in continuous time'

I've not actively worked with any of these though.
Martin Maechler, ETH Zurich



    >> From: s-news-owner@lists.biostat.wustl.edu
    >> [mailto:s-news-owner@lists.biostat.wustl.edu] On
    >> Behalf Of Juan Carlos Ruilova Teran Sent: Wednesday,
    >> May 03, 2006 7:30 AM To: s-news@wubios.wustl.edu
    >> Subject: [S] Hidden Markov Models


    >> Dear All, I'm trying to find code to Hidden Markov
    >> Models (Discrete Time and Continuous Time).  Can
    >> somebody help me with this?  I'm a user of S-Plus 7.0
    >> for Windows.  Thanks Juan Carlos

<Prev in Thread] Current Thread [Next in Thread>