VectorClusterable Class Reference

VectorClusterable wraps vectors in a form accessible to generic clustering algorithms. More...

#include <clusterable-classes.h>

Inheritance diagram for VectorClusterable:

Collaboration diagram for VectorClusterable:

Public Member Functions
	VectorClusterable ()
	VectorClusterable (const Vector< BaseFloat > &vector, BaseFloat weight)
virtual std::string	Type () const
	Return a string that describes the inherited type. More...
virtual BaseFloat	Objf () const
	Return the objective function associated with the stats [assuming ML estimation]. More...
virtual void	SetZero ()
	Set stats to empty. More...
virtual void	Add (const Clusterable &other_in)
	Add other stats. More...
virtual void	Sub (const Clusterable &other_in)
	Subtract other stats. More...
virtual BaseFloat	Normalizer () const
	Return the normalizer (typically, count) associated with the stats. More...
virtual Clusterable *	Copy () const
	Return a copy of this object. More...
virtual void	Scale (BaseFloat f)
	Scale the stats by a positive number f [not mandatory to supply this]. More...
virtual void	Write (std::ostream &os, bool binary) const
	Write data to stream. More...
virtual Clusterable *	ReadNew (std::istream &is, bool binary) const
	Read data from a stream and return the corresponding object (const function; it's a class member because we need access to the vtable so generic code can read derived types). More...
virtual	~VectorClusterable ()
Public Member Functions inherited from Clusterable
virtual	~Clusterable ()
virtual BaseFloat	ObjfPlus (const Clusterable &other) const
	Return the objective function of the combined object this + other. More...
virtual BaseFloat	ObjfMinus (const Clusterable &other) const
	Return the objective function of the subtracted object this - other. More...
virtual BaseFloat	Distance (const Clusterable &other) const
	Return the objective function decrease from merging the two clusters, negated to be a positive number (or zero). More...

Private Member Functions
void	Read (std::istream &is, bool binary)

Private Attributes
double	weight_
Vector< double >	stats_
double	sumsq_

Detailed Description

VectorClusterable wraps vectors in a form accessible to generic clustering algorithms.

Each vector is associated with a weight; these could be 1.0. The objective function (to be maximized) is the negated sum of squared distances from the cluster center to each vector, times that vector's weight.

Definition at line 121 of file clusterable-classes.h.

Constructor & Destructor Documentation

VectorClusterable() [1/2]

VectorClusterable ( )

inline

Definition at line 123 of file clusterable-classes.h.

: weight_(0.0), sumsq_(0.0) {}

VectorClusterable() [2/2]

VectorClusterable	(	const Vector< BaseFloat > &	vector,
		BaseFloat	weight
	)

Definition at line 296 of file clusterable-classes.cc.

References KALDI_ASSERT, VectorBase< Real >::Scale(), VectorClusterable::stats_, VectorClusterable::sumsq_, and kaldi::VecVec().

                                                      :
    weight_(weight), stats_(vector), sumsq_(0.0) {
  stats_.Scale(weight);
  KALDI_ASSERT(weight >= 0.0);
  sumsq_ = VecVec(vector, vector) * weight;
}    

~VectorClusterable()

virtual ~VectorClusterable ( )

inlinevirtual

Definition at line 139 of file clusterable-classes.h.

{}

Member Function Documentation

Add()

void Add ( const Clusterable &  other )

virtual

Add other stats.

Implements Clusterable.

Definition at line 225 of file clusterable-classes.cc.

References KALDI_ASSERT, Clusterable::Type(), and VectorClusterable::weight_.

Referenced by kaldi::TestClusterUtilsVector().

                                                       {
  KALDI_ASSERT(other_in.Type() == "vector");
  const VectorClusterable *other =
      static_cast<const VectorClusterable*>(&other_in);
  weight_ += other->weight_;
  stats_.AddVec(1.0, other->stats_);
  sumsq_ += other->sumsq_;
}

Copy()

Clusterable * Copy ( ) const

virtual

Return a copy of this object.

Implements Clusterable.

Definition at line 255 of file clusterable-classes.cc.

References VectorClusterable::stats_, VectorClusterable::sumsq_, and VectorClusterable::weight_.

                                           {
  VectorClusterable *ans = new VectorClusterable();
  ans->weight_ = weight_;
  ans->sumsq_ = sumsq_;
  ans->stats_ = stats_;
  return ans;
}

Normalizer()

virtual BaseFloat Normalizer ( ) const

inlinevirtual

Return the normalizer (typically, count) associated with the stats.

Implements Clusterable.

Definition at line 134 of file clusterable-classes.h.

References GaussClusterable::Copy(), GaussClusterable::ReadNew(), GaussClusterable::Scale(), and GaussClusterable::Write().

{ return weight_; }

Objf()

BaseFloat Objf ( ) const

virtual

Return the objective function associated with the stats [assuming ML estimation].

Implements Clusterable.

Definition at line 305 of file clusterable-classes.cc.

References KALDI_WARN, VectorClusterable::stats_, VectorClusterable::sumsq_, kaldi::VecVec(), and VectorClusterable::weight_.

Referenced by kaldi::TestClusterUtilsVector().

                                        {
  double direct_sumsq;
  if (weight_ > std::numeric_limits<BaseFloat>::min()) {
    direct_sumsq = VecVec(stats_, stats_) / weight_;
  } else {
    direct_sumsq = 0.0;
  }
  // ans is a negated weighted sum of squared distances; it should not be
  // positive.
  double ans = -(sumsq_ - direct_sumsq); 
  if (ans > 0.0) {
    if (ans > 1.0) {
      KALDI_WARN << "Positive objective function encountered (treating as zero): "
                 << ans;
    }
    ans = 0.0;
  }
  return ans;
}

Read()

void Read ( std::istream &  is, bool  binary  )

private

Definition at line 286 of file clusterable-classes.cc.

References kaldi::ExpectToken(), and kaldi::ReadBasicType().

Referenced by VectorClusterable::ReadNew().

                                                        {
  ExpectToken(is, binary, "VCL");  // magic string.
  ExpectToken(is, binary, "<Weight>");
  ReadBasicType(is, binary, &weight_);
  ExpectToken(is, binary, "<Sumsq>");  
  ReadBasicType(is, binary, &sumsq_);
  ExpectToken(is, binary, "<Stats>");    
  stats_.Read(is, binary);
}

ReadNew()

Clusterable * ReadNew ( std::istream &  os, bool  binary  ) const

virtual

Read data from a stream and return the corresponding object (const function; it's a class member because we need access to the vtable so generic code can read derived types).

Implements Clusterable.

Definition at line 280 of file clusterable-classes.cc.

References VectorClusterable::Read().

                                                                         {
  VectorClusterable *vc = new VectorClusterable();
  vc->Read(is, binary);
  return vc;
}

Scale()

void Scale ( BaseFloat  f )

virtual

Scale the stats by a positive number f [not mandatory to supply this].

Reimplemented from Clusterable.

Definition at line 263 of file clusterable-classes.cc.

References KALDI_ASSERT.

                                         {
  KALDI_ASSERT(f >= 0.0);
  weight_ *= f;
  stats_.Scale(f);
  sumsq_ *= f;
}

SetZero()

virtual void SetZero ( )

inlinevirtual

Set stats to empty.

Implements Clusterable.

Definition at line 131 of file clusterable-classes.h.

References GaussClusterable::Add(), MatrixBase< Real >::Set(), GaussClusterable::stats_, and GaussClusterable::Sub().

{ weight_ = 0.0; sumsq_ = 0.0; stats_.Set(0.0); }

Sub()

void Sub ( const Clusterable &  other )

virtual

Subtract other stats.

Implements Clusterable.

Definition at line 234 of file clusterable-classes.cc.

References KALDI_ASSERT, KALDI_WARN, Clusterable::Type(), and VectorClusterable::weight_.

                                                       {
  KALDI_ASSERT(other_in.Type() == "vector");
  const VectorClusterable *other =
      static_cast<const VectorClusterable*>(&other_in);
  weight_ -= other->weight_;
  sumsq_ -= other->sumsq_;
  stats_.AddVec(-1.0, other->stats_);
  if (weight_ < 0.0) {
    if (weight_ < -0.1 && weight_ < -0.0001 * fabs(other->weight_)) {
      // a negative weight may indicate an algorithmic error if it is
      // encountered.
      KALDI_WARN << "Negative weight encountered " << weight_;
    }
    weight_ = 0.0;
  }
  if (weight_ == 0.0) {
    sumsq_ = 0.0;
    stats_.Set(0.0);
  }
}

Type()

virtual std::string Type ( ) const

inlinevirtual

Return a string that describes the inherited type.

Implements Clusterable.

Definition at line 128 of file clusterable-classes.h.

References GaussClusterable::Objf().

{  return "vector"; }

Write()

void Write ( std::ostream &  os, bool  binary  ) const

virtual

Write data to stream.

Implements Clusterable.

Definition at line 270 of file clusterable-classes.cc.

References kaldi::WriteBasicType(), and kaldi::WriteToken().

                                                               {
  WriteToken(os, binary, "VCL");  // magic string.
  WriteToken(os, binary, "<Weight>");
  WriteBasicType(os, binary, weight_);
  WriteToken(os, binary, "<Sumsq>");  
  WriteBasicType(os, binary, sumsq_);
  WriteToken(os, binary, "<Stats>");    
  stats_.Write(os, binary);
}

Member Data Documentation

stats_

Vector<double> stats_

private

Definition at line 143 of file clusterable-classes.h.

Referenced by VectorClusterable::Copy(), VectorClusterable::Objf(), and VectorClusterable::VectorClusterable().

sumsq_

double sumsq_

private

Definition at line 144 of file clusterable-classes.h.

Referenced by VectorClusterable::Copy(), VectorClusterable::Objf(), and VectorClusterable::VectorClusterable().

weight_

double weight_

private

Definition at line 142 of file clusterable-classes.h.

Referenced by VectorClusterable::Add(), VectorClusterable::Copy(), VectorClusterable::Objf(), and VectorClusterable::Sub().

---

The documentation for this class was generated from the following files:

• tree/clusterable-classes.h
• tree/clusterable-classes.cc