Adding Log F Distribution to runjags module

[Achmad Syahrul Choir]

Dear Forum,
I added a logF distribution into runjags module, but i failed to compile it.

This is my code in DLogf.h

#ifndef DLOGF_H_
#define DLOGF_H_
// Checks the JAGS version and sets necessary macros:
#include "../jagsversions.h"

#ifndef INCLUDERSCALARDIST
#include <distribution/RScalarDist.h>
namespace jags {
#else
#include "jags/RScalarDist.h"
#endif  /* INCLUDERSCALARDIST */

namespace runjags {

/**
 * <pre>
 *
 * </pre>
 * @short Normal distribution
 */
class DLogf : public RScalarDist {
 public:
   DLogf();

  double d(double z, PDFType type,
       std::vector<double const *> const &parameters, 
       bool give_log) const;
  double p(double q, std::vector<double const *> const &parameters, bool lower,
       bool give_log) const;
  double q(double p, std::vector<double const *> const &parameters, bool lower,
       bool log_p) const;
  double r(std::vector<double const *> const &parameters, RNG *rng) const;
  /**
   * Checks that tau > 0
   */
  bool checkParameterValue(std::vector<double const *> const &parameters) const;
  /**
   *    * that is built into the JAGS library, overloading the generic
   * functionality of RScalarDist.
   */

};
}// namespace runjags
#ifndef INCLUDERSCALARDIST
}  // namespace jags
#endif  /* INCLUDERSCALARDIST */
#endif /* DMNLOGF_H_ */

Code in DLogf.cc

#include "DLogf.h"
#include <util/nainf.h>
#include <rng/RNG.h>

#include <cmath>
#include <cfloat>
#include <JRmath.h>
using std::vector;
using std::exp;
using std::log;
using std::pow;

#define ALPHA1(par) (*par[0])
#define ALPHA2(par) (*par[1])

#ifndef INCLUDERSCALARDIST
namespace jags {
#endif  /* INCLUDERSCALARDIST */

namespace runjags {

DLogf::DLogf()
  : RScalarDist("dmnlogf",2, DIST_UNBOUNDED)
{}

bool DLogf::checkParameterValue (vector<double const *> const &par) const
{
  return ( ALPHA1(par) > 0 && ALPHA2(par) > 0 );
}

//Computing pdf

double
DLogf::d(double z, PDFType type,
            vector<double const *> const &par, bool give_log) const
{

  double w=exp(z);
  double log_pdf=dF(w,2*ALPHA1(par),2*ALPHA2(par),1)+z;
  return give_log?log_pdf: exp(log_pdf);
}

//Computing CDF
double
DLogf::p(double z, vector<double const *> const &par, bool lower, bool give_log)
const
{

  double w = exp(z);  
  return pF(w,2*ALPHA1(par),2*ALPHA2(par),lower,give_log);

}

//Computing Invers CDF

double 
DLogf::q(double p, vector<double const *> const &par, bool lower, bool log_p)
const
{
  if ( (log_p  && p > 0) || (!log_p && (p < 0 || p > 1)) )          
    return JAGS_NAN;

  double  w =qF(p,2*ALPHA1(par),2*ALPHA2(par),lower,log_p);
  return log(w);

}

double 
DLogf::r(vector<double const *> const &par, RNG *rng) const
{
  return q(rng->uniform(), par,true, false);  
}

}
#ifndef INCLUDERSCALARDIST
}  // namespace jags
#endif  /* INCLUDERSCALARDIST */

I had a errormassage while i compiled it

distributions/DLogf.o:DLogf.cc:(.text+0x67): undefined reference to jags_dF' distributions/DLogf.o:DLogf.cc:(.text+0x14a): undefined reference tojags_qF'
distributions/DLogf.o:DLogf.cc:(.text+0x1dc): undefined reference to jags_qF' distributions/DLogf.o:DLogf.cc:(.text+0xd7): undefined reference tojags_pF'

How to solve this problem? Thank You.

Best Regard

Achmad Syahrul

[Achmad Syahrul Choir]

I have successfully fixed the code so it can be compiled. I convert JRmath to Rmath, dF to df, pF to pf and qF to qf. However, when I use this distribution, runjags (JAGS) do not report the DIC value. Why JAGS can't calculate DIC for this distribution? Thank you.

[Martyn Plummer]

That is not the way to solve your first problem. You need to add the library jrmath, which provides the missing symbols jags_qF, jags_pF, jags_dF when you compile the shared object runjags.so by modifying the file src/Makevars.in. The line

 PKG_LIBS=-L@JAGS_LIB@ -ljags @JAGS_RPATH@

should be changed to

 PKG_LIBS=-L@JAGS_LIB@ -ljags -ljrmath @JAGS_RPATH@

and you will need to modify Makevars.win if you want to compile your modified package on Windows.

[Achmad Syahrul Choir]

My problem was solved by your suggestion. Thank you.