PLearn 0.1
Public Types | Public Member Functions | Static Public Member Functions | Public Attributes | Static Public Attributes | Static Protected Member Functions | Private Member Functions
PLearn::ScaledConditionalCDFSmoother Class Reference

#include <ScaledConditionalCDFSmoother.h>

Inheritance diagram for PLearn::ScaledConditionalCDFSmoother:
Inheritance graph
[legend]
Collaboration diagram for PLearn::ScaledConditionalCDFSmoother:
Collaboration graph
[legend]

List of all members.

Public Types

typedef ConditionalCDFSmoother inherited

Public Member Functions

 ScaledConditionalCDFSmoother ()
virtual void build ()
 Post-constructor.
virtual void makeDeepCopyFromShallowCopy (CopiesMap &copies)
 Transforms a shallow copy into a deep copy.
virtual string classname () const
virtual OptionListgetOptionList () const
virtual OptionMapgetOptionMap () const
virtual RemoteMethodMapgetRemoteMethodMap () const
virtual
ScaledConditionalCDFSmoother
deepCopy (CopiesMap &copies) const
virtual real smooth (const Vec &source_function, Vec &smoothed_function, Vec bin_positions=Vec(), Vec dest_bin_positions=Vec()) const

Static Public Member Functions

static string _classname_ ()
 Declares name and deepCopy methods.
static OptionList_getOptionList_ ()
static RemoteMethodMap_getRemoteMethodMap_ ()
static Object_new_instance_for_typemap_ ()
static bool _isa_ (const Object *o)
static void _static_initialize_ ()
static const PPathdeclaringFile ()

Public Attributes

bool preserve_relative_density

Static Public Attributes

static StaticInitializer _static_initializer_

Static Protected Member Functions

static void declareOptions (OptionList &ol)
 Declares this class' options.

Private Member Functions

void build_ ()
 This does the actual building.

Detailed Description

Definition at line 52 of file ScaledConditionalCDFSmoother.h.


Member Typedef Documentation

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 64 of file ScaledConditionalCDFSmoother.h.


Constructor & Destructor Documentation

PLearn::ScaledConditionalCDFSmoother::ScaledConditionalCDFSmoother ( )

Member Function Documentation

string PLearn::ScaledConditionalCDFSmoother::_classname_ ( ) [static]

Declares name and deepCopy methods.

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 65 of file ScaledConditionalCDFSmoother.cc.

OptionList & PLearn::ScaledConditionalCDFSmoother::_getOptionList_ ( ) [static]

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 65 of file ScaledConditionalCDFSmoother.cc.

RemoteMethodMap & PLearn::ScaledConditionalCDFSmoother::_getRemoteMethodMap_ ( ) [static]

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 65 of file ScaledConditionalCDFSmoother.cc.

bool PLearn::ScaledConditionalCDFSmoother::_isa_ ( const Object o) [static]

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 65 of file ScaledConditionalCDFSmoother.cc.

Object * PLearn::ScaledConditionalCDFSmoother::_new_instance_for_typemap_ ( ) [static]

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 65 of file ScaledConditionalCDFSmoother.cc.

StaticInitializer ScaledConditionalCDFSmoother::_static_initializer_ & PLearn::ScaledConditionalCDFSmoother::_static_initialize_ ( ) [static]

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 65 of file ScaledConditionalCDFSmoother.cc.

void PLearn::ScaledConditionalCDFSmoother::build ( ) [virtual]

Post-constructor.

The normal implementation should call simply inherited::build(), then this class's build_(). This method should be callable again at later times, after modifying some option fields to change the "architecture" of the object.

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 94 of file ScaledConditionalCDFSmoother.cc.

References PLearn::ConditionalCDFSmoother::build(), and build_().

Here is the call graph for this function:

void PLearn::ScaledConditionalCDFSmoother::build_ ( ) [private]

This does the actual building.

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 89 of file ScaledConditionalCDFSmoother.cc.

Referenced by build().

{
}

Here is the caller graph for this function:

string PLearn::ScaledConditionalCDFSmoother::classname ( ) const [virtual]

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 65 of file ScaledConditionalCDFSmoother.cc.

void PLearn::ScaledConditionalCDFSmoother::declareOptions ( OptionList ol) [static, protected]

Declares this class' options.

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 67 of file ScaledConditionalCDFSmoother.cc.

References PLearn::OptionBase::buildoption, PLearn::declareOption(), PLearn::ConditionalCDFSmoother::declareOptions(), and preserve_relative_density.

{
    declareOption(ol, "preserve_relative_density", &ScaledConditionalCDFSmoother::preserve_relative_density, 
                  OptionBase::buildoption,
                  "If true then the following formula is used inside each of the large intervals (t_0,t_1):\n"
                  "  S(y_t) = S(y_{t_0})+(PS(y_t)-PS(y_{t_0}))(RS(y_{t_0})-RS(y_{t_1}))/(PS(y_{t_1})-PS(y_{t_0})\n"
                  "where S(y_t) is the smoothed survival function at position y_t, PS(y_t) is the prior\n"
                  "survival function at y_t, and RS(y_t) is the rough survival function (which is to be\n"
                  "smoothed) at y_t. Note that RS is only known at the extremes of the interval, y_{t_0}\n"
                  "and y_{t_1}. Note that this formula has the property that within the interval, the\n"
                  "density is the prior density, scaled by the ratio of the total density in the interval\n"
                  "for the target rough curve with respect to the prior curve\n"
                  "If false, then the following formula is used instead, using the same notation:\n"
                  "  S(y_t) = PS(y_t)(RS(y_{t_0})/PS(y_{t_0}) + (y_t - y_{t_0})(RS(y_{t_1})-RS(y_{t_0}))/(PS(y_{t_1}) (t_1 - t_0)))\n"
                  "What is the justification for this second formula?\n"
        );
                

    // Now call the parent class' declareOptions
    inherited::declareOptions(ol);
}

Here is the call graph for this function:

static const PPath& PLearn::ScaledConditionalCDFSmoother::declaringFile ( ) [inline, static]

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 106 of file ScaledConditionalCDFSmoother.h.

:
    // The source function is either f(i) = source_function[i] as a function of i
ScaledConditionalCDFSmoother * PLearn::ScaledConditionalCDFSmoother::deepCopy ( CopiesMap copies) const [virtual]

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 65 of file ScaledConditionalCDFSmoother.cc.

OptionList & PLearn::ScaledConditionalCDFSmoother::getOptionList ( ) const [virtual]

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 65 of file ScaledConditionalCDFSmoother.cc.

OptionMap & PLearn::ScaledConditionalCDFSmoother::getOptionMap ( ) const [virtual]

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 65 of file ScaledConditionalCDFSmoother.cc.

RemoteMethodMap & PLearn::ScaledConditionalCDFSmoother::getRemoteMethodMap ( ) const [virtual]

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 65 of file ScaledConditionalCDFSmoother.cc.

void PLearn::ScaledConditionalCDFSmoother::makeDeepCopyFromShallowCopy ( CopiesMap copies) [virtual]

Transforms a shallow copy into a deep copy.

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 101 of file ScaledConditionalCDFSmoother.cc.

References PLearn::ConditionalCDFSmoother::makeDeepCopyFromShallowCopy().

Here is the call graph for this function:

real PLearn::ScaledConditionalCDFSmoother::smooth ( const Vec source_function,
Vec smoothed_function,
Vec  bin_positions = Vec(),
Vec  dest_bin_positions = Vec() 
) const [virtual]

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 111 of file ScaledConditionalCDFSmoother.cc.

References PLearn::endl(), i, j, PLERROR, preserve_relative_density, PLearn::ConditionalCDFSmoother::prior_cdf, PLearn::TVec< T >::resize(), and PLearn::TVec< T >::size().

{
    // put in 'survival_fn' the multiplicatively adjusted unconditional_survival_fn
    // such that the estimatedS values at yvalues match. In each segment
    // between prev_y and next_y. The adjustment ratio varies linearly from
    // estimatedS[prev_y]/unconditionalS[prev_y] to estimatedS[next_y]/unconditionalS[next_y]):
    //  prev_ratio = estimatedS[prev_y]/unconditionalS[prev_y]
    //  next_ratio = estimatedS[next_y]/unconditionalS[next_y]
    //  adjustment = prev_ratio + (y-prev_y)*next_ratio/(next_y-prev_y)
    //  s(y) = unconditional_s(y)*adjustment

    if (!prior_cdf)
        PLERROR("in ScaledConditionalCDFSmoother::smooth  you need to supply a prior_cdf");
    //assume source_function is a survival fn.
    if(bin_positions.size() != source_function.size()+1)
        PLERROR("in ScaledConditionalCDFSmoother::smooth  you need to supply bin_positions");
    if(dest_bin_positions.size() == 0)
        PLERROR("in ScaledConditionalCDFSmoother::smooth  you need to supply dest_bin_positions");
    smoothed_function.resize(dest_bin_positions.size()-1);
 

    int j= 0;
    for(int i= 0; i < source_function.size(); ++i)
    {
        Vec v0(1), v1(1);//prev_y, next_y
        v0[0]= bin_positions[i];
        v1[0]= bin_positions[i+1];
      
        real prev_ratio=  source_function[i]/prior_cdf->survival_fn(v0);
        real next_ratio;
        if(i == source_function.size()-1)
            next_ratio= 0.0;
        else
            next_ratio=  source_function[i+1]/prior_cdf->survival_fn(v1);
      
        cout  << source_function[i] << '\t'  << prev_ratio  << '\t' << next_ratio << '\t' << v0[0] << '\t' << v1[0] << endl;
        real slope = !preserve_relative_density? 0 : 
            ((source_function[i+1]-source_function[i])/(prior_cdf->survival_fn(v1)-prior_cdf->survival_fn(v0)));
        real absisse = !preserve_relative_density? 0 : 
            (source_function[i] - slope * prior_cdf->survival_fn(v0));
        while(j < smoothed_function.size() && dest_bin_positions[j+1] <= bin_positions[i+1])
        {
            Vec v(1);
            v[0]= dest_bin_positions[j];
            // the line below seems wrong, so I have fixed it -- YB
            // the reason it seems wrong is that smoothed_function[j_final] should be equal
            // to source_function[i+1], but it is not, currently.
            // smoothed_function[j]= prior_cdf->survival_fn(v) * (prev_ratio + (v[0]-v0[0])*next_ratio/(v1[0]-v0[0]));
            if (!preserve_relative_density)
                smoothed_function[j]= prior_cdf->survival_fn(v) * 
                    (prev_ratio + (v[0]-v0[0])*(next_ratio-prev_ratio)/(v1[0]-v0[0]));
            else // scale with bin number, i.e. warped with density
                smoothed_function[j]= absisse + slope * prior_cdf->survival_fn(v);
            cout  << '\t' << v[0] << '\t' << prior_cdf->survival_fn(v) << '\t' << smoothed_function[j] << endl;
            ++j;
        }
    }
  
  












    /*
    //assume source_function is a survival fn.
    if(bin_positions.size() != source_function.size()+1)
    PLERROR("in ScaledConditionalCDFSmoother::smooth  you need to supply bin_positions");
    if(dest_bin_positions.size() == 0)
    PLERROR("in ScaledConditionalCDFSmoother::smooth  you need to supply dest_bin_positions");
    smoothed_function.resize(dest_bin_positions.size()-1);
    Vec f0(dest_bin_positions.size()-1); //new density


    int j= 0;
    real factor= 1.0;
    for(int i= 0; i < source_function.size(); ++i)
    {
    Vec v0(1), v1(1);
    v0[0]= bin_positions[i];
    v1[0]= bin_positions[i+1];
    real prior_prob= prior_cdf->survival_fn(v0) - prior_cdf->survival_fn(v1);
    real prob;
    if(i < source_function.size()-1)
    prob= (source_function[i]-source_function[i+1]);
    else
    prob= source_function[i];

    if(0 < prior_prob && prob != 0.0)
    factor= prob / prior_prob;
    // else: use prev. factor

    //dummy-temp  
    cout << v0[0] << '-' << v1[0] << ":\t" << prob << '/' <<  prior_prob << '=' << factor << endl;


    while(j < smoothed_function.size() && dest_bin_positions[j+1] <= bin_positions[i+1])
    {
    Vec v(1);
    v[0]= (dest_bin_positions[j]+dest_bin_positions[j+1])/2;
    //    smoothed_function[j]= factor * prior_cdf->survival_fn(v);
    f0[j]= factor * prior_cdf->density(v);
    //dummy-temp  
    cout << '\t' << smoothed_function[j] << "= " <<  factor  << " * " << prior_cdf->survival_fn(v) << endl;

    ++j;
    }
    }
  

    HistogramDistribution::calc_survival_from_density(f0, smoothed_function, dest_bin_positions);

    */

    /*
      int j= 0;
      real factor= 1.0;
      for(int i= 0; i < source_function.size(); ++i)
      {
      Vec v0(1), v1(1);
      v0[0]= bin_positions[i];
      v1[0]= bin_positions[i+1];
      real prior_prob= prior_cdf->survival_fn(v0) - prior_cdf->survival_fn(v1);
      real prob;
      if(i < source_function.size()-1)
      prob= (source_function[i]-source_function[i+1]);
      else
      prob= source_function[i];

      if(0 < prior_prob && prob != 0.0)
      factor= prob / prior_prob;
      // else: use prev. factor

      //dummy-temp  
      cout << v0[0] << '-' << v1[0] << ":\t" << prob << '/' <<  prior_prob << '=' << factor << endl;


      while(j < smoothed_function.size() && dest_bin_positions[j+1] <= bin_positions[i+1])
      {
      Vec v(1);
      v[0]= (dest_bin_positions[j]+dest_bin_positions[j+1])/2;
      smoothed_function[j]= factor * prior_cdf->survival_fn(v);
      //dummy-temp  
      cout << '\t' << smoothed_function[j] << "= " <<  factor  << " * " << prior_cdf->survival_fn(v) << endl;

      ++j;
      }
      }
    */

    return 0.0; //dummy - FIXME - xsm
}

Here is the call graph for this function:


Member Data Documentation

Reimplemented from PLearn::ConditionalCDFSmoother.

Definition at line 106 of file ScaledConditionalCDFSmoother.h.

Definition at line 74 of file ScaledConditionalCDFSmoother.h.

Referenced by declareOptions(), and smooth().


The documentation for this class was generated from the following files:
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Defines