SimpleDiscreteDistribution.cpp

Go to the documentation of this file.

//
// File: SimpleDiscreteDistribution.cpp
// Created by: Julien Dutheil
// Created on: ?
//

/*
   Copyright or © or Copr. Bio++ Development Team, (November 17, 2004)

   This software is a computer program whose purpose is to provide classes
   for numerical calculus.

   This software is governed by the CeCILL  license under French law and
   abiding by the rules of distribution of free software.  You can  use,
   modify and/ or redistribute the software under the terms of the CeCILL
   license as circulated by CEA, CNRS and INRIA at the following URL
   "http://www.cecill.info".

   As a counterpart to the access to the source code and  rights to copy,
   modify and redistribute granted by the license, users are provided only
   with a limited warranty  and the software's author,  the holder of the
   economic rights,  and the successive licensors  have only  limited
   liability.

   In this respect, the user's attention is drawn to the risks associated
   with loading,  using,  modifying and/or developing or reproducing the
   software by the user in light of its specific status of free software,
   that may mean  that it is complicated to manipulate,  and  that  also
   therefore means  that it is reserved for developers  and  experienced
   professionals having in-depth computer knowledge. Users are therefore
   encouraged to load and test the software's suitability as regards their
   requirements in conditions enabling the security of their systems and/or
   data to be ensured and,  more generally, to use and operate it in the
   same conditions as regards security.

   The fact that you are presently reading this means that you have had
   knowledge of the CeCILL license and that you accept its terms.
 */

#include "SimpleDiscreteDistribution.h"
#include "../NumConstants.h"
#include "../../Utils/MapTools.h"
#include "../../Text/TextTools.h"

using namespace bpp;
using namespace std;


SimpleDiscreteDistribution::SimpleDiscreteDistribution(const map<double, double>& distribution,
                                                       double prec,
                                                       bool fixed) :
  AbstractParameterAliasable("Simple."),
  AbstractDiscreteDistribution(distribution.size(), prec, "Simple."),
  givenRanges_()
{
  double sum = 0;
  for (map<double, double>::const_iterator i = distribution.begin(); i != distribution.end(); i++)
  {
    distribution_[i->first] = i->second;
    sum += i->second;
  }
  if (fabs(1. - sum) > precision())
    throw Exception("SimpleDiscreteDistribution. Probabilities must equal 1 (sum =" + TextTools::toString(sum) + ").");

  if (!fixed)
  {
    unsigned int n = 1;
    double x, y = 1;
    for (map<double, double>::const_iterator i = distribution.begin(); i != distribution.end(); i++)
    {
      addParameter_(new Parameter("Simple.V" + TextTools::toString(n), i->first));

      if (n != numberOfCategories_)
      {
        x = i->second;
        addParameter_(new Parameter("Simple.theta" + TextTools::toString(n), x / y, &Parameter::PROP_CONSTRAINT_IN));
        y -= x;
      }
      n++;
    }
  }
  discretize();
}

SimpleDiscreteDistribution::SimpleDiscreteDistribution(const vector<double>& values,
                                                       const vector<double>& probas,
                                                       double prec,
                                                       bool fixed
                                                       ) :
  AbstractParameterAliasable("Simple."),
  AbstractDiscreteDistribution(values.size(), prec, "Simple."),
  givenRanges_()
{
  if (values.size() != probas.size())
  {
    throw Exception("SimpleDiscreteDistribution. Values and probabilities vectors must have the same size (" + TextTools::toString(values.size()) + " != " + TextTools::toString(probas.size()) + ").");
  }
  size_t size = values.size();

  for (size_t i = 0; i < size; i++)
  {
    if (distribution_.find(values[i]) != distribution_.end())
      throw Exception("SimpleDiscreteDistribution: two given values are equal");
    else
      distribution_[values[i]] = probas[i];
  }

  double sum = VectorTools::sum(probas);
  if (fabs(1. - sum) > precision())
    throw Exception("SimpleDiscreteDistribution. Probabilities must equal 1 (sum =" + TextTools::toString(sum) + ").");

  if (!fixed)
  {
    double y = 1;
    for (unsigned int i = 0; i < size - 1; i++)
    {
      addParameter_(new Parameter("Simple.V" + TextTools::toString(i + 1), values[i]));
      addParameter_(new Parameter("Simple.theta" + TextTools::toString(i + 1), probas[i] / y, &Parameter::PROP_CONSTRAINT_IN));
      y -= probas[i];
    }
    addParameter_(new Parameter("Simple.V" + TextTools::toString(size), values[size - 1]));
  }

  discretize();
}

SimpleDiscreteDistribution::SimpleDiscreteDistribution(const std::vector<double>& values,
                                                       const std::map<size_t, std::vector<double> >& ranges,
                                                       const std::vector<double>& probas,
                                                       double prec,
                                                       bool fixed) :
  AbstractParameterAliasable("Simple."),
  AbstractDiscreteDistribution(values.size(), prec, "Simple."),
  givenRanges_()
{
  if (values.size() != probas.size())
  {
    throw Exception("SimpleDiscreteDistribution. Values and probabilities vectors must have the same size (" + TextTools::toString(values.size()) + " != " + TextTools::toString(probas.size()) + ").");
  }
  size_t size = values.size();

  for (size_t i = 0; i < size; i++)
  {
    if (distribution_.find(values[i]) != distribution_.end())
      throw Exception("SimpleDiscreteDistribution: two given values are equal");
    else
      distribution_[values[i]] = probas[i];
  }

  double sum = VectorTools::sum(probas);
  if (fabs(1. - sum) > precision())
    throw Exception("SimpleDiscreteDistribution. Probabilities must equal 1 (sum =" + TextTools::toString(sum) + ").");

  if (!fixed)
  {
    double y = 1;
    for (size_t i = 0; i < size - 1; i++)
    {
      map<size_t, vector<double> >::const_iterator it = ranges.find(i + 1);
      if (it == ranges.end())
        addParameter_(new Parameter("Simple.V" + TextTools::toString(i + 1), values[i]));
      else
      {
        if (values[i] >= it->second[0] &&  values[i] <= it->second[1])
        {
          addParameter_(new Parameter("Simple.V" + TextTools::toString(i + 1), values[i], new IntervalConstraint(it->second[0], it->second[1], true, true), true));
          givenRanges_[i + 1] = it->second;
        }
        else
          throw Exception("SimpleDiscreteDistribution. Value and given range of parameter V" + TextTools::toString(i + 1) + " do not match: " + TextTools::toString(values[i]) + " vs [" + TextTools::toString(it->second[0]) + ";" + TextTools::toString(it->second[1]) + "]");
      }
      addParameter_(new Parameter("Simple.theta" + TextTools::toString(i + 1), probas[i] / y, &Parameter::PROP_CONSTRAINT_IN));
      y -= probas[i];
    }

    map<size_t, vector<double> >::const_iterator it = ranges.find(size);
    if (it == ranges.end())
      addParameter_(new Parameter("Simple.V" + TextTools::toString(size), values[size - 1]));
    else
    {
      if (values[size - 1] >= it->second[0] &&  values[size - 1] <= it->second[1])
      {
        addParameter_(new Parameter("Simple.V" + TextTools::toString(size), values[size - 1], new IntervalConstraint(it->second[0], it->second[1], true, true), true));
        givenRanges_[size] = it->second;
      }
      else
        throw Exception("SimpleDiscreteDistribution. Value and given range of parameter V" + TextTools::toString(size) + " do not match: " + TextTools::toString(values[size - 1]) + " vs [" + TextTools::toString(it->second[0]) + ";" + TextTools::toString(it->second[1]) + "]");
    }
  }

  discretize();
}


SimpleDiscreteDistribution::SimpleDiscreteDistribution(const SimpleDiscreteDistribution& sdd) :
  AbstractParameterAliasable(sdd),
  AbstractDiscreteDistribution(sdd),
  givenRanges_(sdd.givenRanges_)
{}

SimpleDiscreteDistribution& SimpleDiscreteDistribution::operator=(const SimpleDiscreteDistribution& sdd)
{
  AbstractParameterAliasable::operator=(sdd);
  AbstractDiscreteDistribution::operator=(sdd);
  givenRanges_ = sdd.givenRanges_;

  return *this;
}

void SimpleDiscreteDistribution::fireParameterChanged(const ParameterList& parameters)
{
  if (getNumberOfParameters() != 0)
  {
    AbstractDiscreteDistribution::fireParameterChanged(parameters);
    size_t size = distribution_.size();

    distribution_.clear();
    double x = 1.0;
    double v;
    for (size_t i = 0; i < size - 1; i++)
    {
      v = getParameterValue("V" + TextTools::toString(i + 1));
      if (distribution_.find(v) != distribution_.end())
      {
        int j = 1;
        int f = ((v + precision()) >= intMinMax_.getUpperBound()) ? -1 : 1;
        while (distribution_.find(v + f * j * precision()) != distribution_.end())
        {
          j++;
          f = ((v + f * j * precision()) >= intMinMax_.getUpperBound()) ? -1 : 1;
        }
        v += f * j * precision();
        // approximation to avoid useless computings:
        // setParameterValue("V"+TextTools::toString(i+1),v);
      }
      distribution_[v] = getParameterValue("theta" + TextTools::toString(i + 1)) * x;
      x *= 1 - getParameterValue("theta" + TextTools::toString(i + 1));
    }

    v = getParameterValue("V" + TextTools::toString(size));
    if (distribution_.find(v) != distribution_.end())
    {
      int j = 1;
      int f = ((v + precision()) >= intMinMax_.getUpperBound()) ? -1 : 1;
      while (distribution_.find(v + f * j * precision()) != distribution_.end())
      {
        j++;
        f = ((v + f * j * precision()) >= intMinMax_.getUpperBound()) ? -1 : 1;
      }
      v += f * j * precision();
      // approximation to avoid useless computings:
      // setParameterValue("V"+TextTools::toString(size),v);
    }
    distribution_[v] = x;
  }
  discretize();
}

double SimpleDiscreteDistribution::qProb(double x) const
{
  double s = -NumConstants::VERY_BIG();
  double x2 = x;
  for (map<double, double>::const_iterator it = distribution_.begin(); it != distribution_.end(); it++)
  {
    x2 -= it->second;
    if (x2 < 0)
      return s;
    else
      s = it->second;
  }

  return s;
}

double SimpleDiscreteDistribution::pProb(double x) const
{
  double s = 0;
  for (map<double, double>::const_iterator it = distribution_.begin(); it != distribution_.end(); it++)
  {
    if (it->first >= x)
      s += it->second;
    else
      break;
  }

  return s;
}

double SimpleDiscreteDistribution::Expectation(double a) const
{
  double s = 0;
  for (map<double, double>::const_iterator it = distribution_.begin(); it != distribution_.end(); it++)
  {
    if (it->first >= a)
      s += it->second;
    else
      break;
  }

  return s;
}


void SimpleDiscreteDistribution::discretize()
{
  // Compute a new arbitray bounderi:
  vector<double> values = MapTools::getKeys<double, double, AbstractDiscreteDistribution::Order>(distribution_);

  // Fill from 0 to numberOfCategories_-2 with midpoints:
  for (unsigned int i = 0; i < numberOfCategories_ - 1; i++)
  {
    bounds_[i] = (values[i] + values[i + 1]) / 2.;
  }
}

void SimpleDiscreteDistribution::restrictToConstraint(const Constraint& c)
{
  if (getNumberOfParameters() == 0)
    return;

  const IntervalConstraint* pi = dynamic_cast<const IntervalConstraint*>(&c);

  if (!pi)
    throw Exception("SimpleDiscreteDistribution::restrictToConstraint: Non-interval exception");

  map<double, double>::const_iterator it;

  for (it = distribution_.begin(); it != distribution_.end(); it++)
  {
    if (!pi->isCorrect(it->first))
      throw Exception("Impossible to restrict to Constraint value " + TextTools::toString(it->first));
  }

  AbstractDiscreteDistribution::restrictToConstraint(c);

  size_t size = distribution_.size();
  for (size_t i = 0; i < size; i++)
  {
    map<size_t, vector<double> >::const_iterator itr = givenRanges_.find(i + 1);
    if (itr == givenRanges_.end())
      getParameter_("V" + TextTools::toString(i + 1)).setConstraint(intMinMax_.clone(), true);
    else
    {
      const Constraint* pc = getParameter_("V" + TextTools::toString(i + 1)).removeConstraint();
      getParameter_("V" + TextTools::toString(i + 1)).setConstraint(*pc & *intMinMax_.clone(), true);
      delete pc;
    }
  }
}