getwgt/getwgt.html

<html>

  <head>
    <title>
      GETWGT - Dirichlet Mixture Estimation
    </title>
  </head>

  <body bgcolor="#EEEEEE" link="#CC0000" alink="#FF3300" vlink="#000055">

    <h1 align = "center">
      GETWGT <br> Dirichlet Mixture Estimation
    </h1>

    <p>
      <b>GETWGT</b>
      is a FORTRAN90 library which
      handles Dirichlet Mixture estimation.
    </p>

    <p>
      The main requirement in the design of this library was that there be
      a single routine with a simple interface through which the user
      interacts.  This routine, called <b>getwgt</b>, accepts a set of
      nucleic acid counts, updates an internal Dirichlet mixture model,
      and returns its current estimate for the pseudocounts.  All other
      transactions and information are hidden from the user.
    </p>

    <p>
      The program requires a data file, containing the parameters of the
      Dirichlet density distributions that make up the mixture.
      This file needs to be named <b>weights.txt</b>.  Two files are
      provided here as possible sources of this data.  These files
      were obtained from
      <a href = "http://www.cse.ucsc.edu/research/compbio/dirichlets/">
      the UCSC computational biology page</a>,
      where more information, papers, and data is available.
    </p>

    <h3 align = "center">
      Licensing:
    </h3>

    <p>
      The computer code and data files described and made available on this web page
      are distributed under
      <a href = "../../txt/gnu_lgpl.txt">the GNU LGPL license.</a>
    </p>

    <h3 align = "center">
      Reference:
    </h3>

    <p>
      <ol>
        <li>
          William Cody, Kenneth Hillstrom,<br>
          Chebyshev Approximations for the Natural Logarithm of the
          Gamma Function,
          Mathematics of Computation,<br>
          Volume 21, Number 98, April 1967, pages 198-203.
        </li>
        <li>
          Brian Everitt, David Hand,<br>
          Finite Mixture Distributions,<br>
          Chapman and Hall, 1981.
        </li>
        <li>
          Kenneth Lange,<br>
          Mathematical and Statistical Methods for Genetic Analysis,<br>
          Springer, 1997,<br>
          ISBN: 0387953892,<br>
          LC: QH438.4.M33.L36.
        </li>
        <li>
          AFM Smith, Udi Makov,<br>
          A Quasi-Bayes Sequential Procedure for Mixtures,<br>
          Journal of the Royal Statistical Society,<br>
          Volume 40, Number 1, B, 1978, pages 106-112.
        </li>
      </ol>
    </p>

    <h3 align = "center">
      Source code:
    </h3>

    <p>
      <ul>
        <li>
          <A href = "getwgt.f90">getwgt.f90</A>, the source code;
        </li>
        <li>
          <A href = "getwgt.sh">getwgt.sh</A>, the source code;
        </li>
      </ul>
    </p>

    <h3 align = "center">
      Examples and Tests:
    </h3>

    <p>
      <ul>
        <li>
          <A href = "getwgt_prb.f90">getwgt_prb.f90</A>, a sample
          calling program;
        </li>
        <li>
          <A href = "getwgt_prb.sh">getwgt_prb.sh</A>,
          commands to compile, link and run the sample calling program;
        </li>
        <li>
          <A href = "getwgt_prb_output.txt">getwgt_prb_output.txt</A>, the output from the
          sample program;
        </li>
        <li>
          <A href = "byst_weights.txt">byst_weights.txt</A>, the "current best"
          data defining a Dirichlet mixture.
        </li>
        <li>
          <A href = "blocks_weights.txt">blocks_weights.txt</A>, the BLOCKS
          data defining a Dirichlet mixture.
        </li>
      </ul>
    </p>

    <h3 align = "center">
      List of Routines:
    </h3>

    <p>
      <ul>
        <li>
          <b>GETWGT</b> updates the Dirichlet mixture weights based on a set of counts.
        </li>
        <li>
          <b>CH_CAP</b> capitalizes a single character.
        </li>
        <li>
          <b>CH_EQI</b> is a case insensitive comparison of two characters for equality.
        </li>
        <li>
          <b>CH_NEXT</b> "reads" space-separated characters from a string, one at a time.
        </li>
        <li>
          <b>CH_TO_DIGIT</b> returns the integer value of a base 10 digit.
        </li>
        <li>
          <b>COMP_PARAM_PRINT</b> prints the parameters for the mixture components.
        </li>
        <li>
          <b>DIRICHLET_MEAN</b> returns the means of the Dirichlet PDF.
        </li>
        <li>
          <b>DIRICHLET_MULTINOMIAL_PDF</b> evaluates a Dirichlet Multinomial PDF.
        </li>
        <li>
          <b>EVENT_PROCESS</b> updates the mixture weight distribution parameters.
        </li>
        <li>
          <b>GAMMA_LOG</b> calculates the natural logarithm of GAMMA ( X ) for positive X.
        </li>
        <li>
          <b>I4_NEXT</b> "reads" integers from a string, one at a time.
        </li>
        <li>
          <b>MIXTURE_READ</b> reads the Dirichlet mixture parameters from a file.
        </li>
        <li>
          <b>R4_NEXT</b> "reads" real numbers from a string, one at a time.
        </li>
        <li>
          <b>R4VEC_COPY</b> copies a real vector.
        </li>
        <li>
          <b>R4VEC_PRINT</b> prints a real vector, with an optional title.
        </li>
        <li>
          <b>R4VEC_UNIT_SUM</b> normalizes a real vector to have unit sum.
        </li>
        <li>
          <b>S_BEGIN</b> is TRUE if one string matches the beginning of the other.
        </li>
        <li>
          <b>S_TO_I4</b> reads an integer value from a string.
        </li>
        <li>
          <b>S_TO_R4</b> reads a real number from a string.
        </li>
      </ul>
    </p>

    <p>
      You can go up one level to <a href = "../f_src.html">
      the FORTRAN90 source codes</a>.
    </p>

    <hr>

    <i>
      Last revised on 30 December 2007.
    </i>

    <!-- John Burkardt -->

  </body>

</html>