dataset Class Template Reference

Class template for storing, retrieving, and manipulating a vector of input-output style data. More...

#include <dataset.h>

Collaboration diagram for dataset:

[legend]List of all members.

Public Types
typedef Tx	x_type
typedef Ty	y_type
Public Member Functions
	dataset ()
dataset &	operator+= (const dataset &ds)
	Combine two data sets.
void	append (const Tx &_x, const Ty &_y)
void	replace (UINT i, const Tx &_x, const Ty &_y)
UINT	size () const
bool	empty () const
void	clear ()
const Tx &	x (UINT i) const
const Ty &	y (UINT i) const
template<typename IIX, typename IIY>
	dataset (IIX xb, IIX xe, IIY yb, IIY ye)
template<typename IIX, typename IIY>
void	import (IIX xb, IIX xe, IIY yb, IIY ye)
	Import data from other types.
dataset *	random_sample (UINT n) const
	Generate a randomly sampled copy of the data set.
template<typename W>
dataset *	random_sample (const W &wgt, UINT n) const

---

Detailed Description

template<typename Tx, typename Ty>
class lemga::dataset< Tx, Ty >

Class template for storing, retrieving, and manipulating a vector of input-output style data.

Parameters:

	Tx	Type of input x.
	Ty	Type of output y.

What we want are: 1. a place to store data 2. be able to deal with missing feature?

Member functions:

• basic info.: size, dimension, ...
• load from / save to a file
• copy from / to another format (REAL **)
• ways to produce new data set (by sampling...)

Todo:

documentation

Definition at line 36 of file dataset.h.

---

Member Typedef Documentation

typedef Tx x_type

Definition at line 47 of file dataset.h.

typedef Ty y_type

Definition at line 48 of file dataset.h.

---

Constructor & Destructor Documentation

dataset (   )  [inline]

Definition at line 50 of file dataset.h.

dataset (  IIX  xb, IIX  xe, IIY  yb, IIY  ye )  [inline]

Definition at line 62 of file dataset.h.

---

Member Function Documentation

void append (  const Tx &  _x, const Ty &  _y )  [inline]

Definition at line 156 of file dataset.h. Referenced by HoldoutCrossVal::cv_round(), vFoldCrossVal::cv_round(), Ordinal_BLE::extend_data(), lemga::load_data(), DataFeeder::next_train_test(), AdaBoost_ECOC::train_with_full_partition(), AdaBoost_ERP::train_with_partial_partition(), and MultiClass_ECOC::train_with_partition().

void clear (   )  [inline]

Definition at line 55 of file dataset.h.

bool empty (   )  const [inline]

Definition at line 54 of file dataset.h.

void import (  IIX  xb, IIX  xe, IIY  yb, IIY  ye )  [inline]

Import data from other types. Parameters: IIX  stands for Input Iterator for x. IIY  stands for Input Iterator for y. import() copies input x from range [xb, xe) and output y from range [yb, ye). Old data stored in the set will be erased. To import data from two vectors vx and vy, use import(vx.begin(), vx.end(), vy.begin(), vy.end()); To import n samples from two pointers px and py, use import(px, px+n, py, py+n); Definition at line 78 of file dataset.h.

dataset& operator+= (  const dataset< Tx, Ty > &  ds  )  [inline]

Combine two data sets. Note: Code copy(ds.d.begin(), ds.d.end(), back_inserter(d)); does almost the same thing, but will fail when ds is just *this. That is, ds += ds; doesn't work. Todo: We need more functions to add/remove samples. Definition at line 148 of file dataset.h. References dataset::d.

dataset* random_sample (  const W &  wgt, UINT  n )  const [inline]

Generate a randomly sampled copy of the data set. Parameters: n  Number of random samples requested. Returns: A pointer to the new born data set. Samples are chosen with uniform probability. Parameters: W  Sample weight type, which usually is vector<REAL>. W should support operator[]. wgt  Sample weight. Definition at line 113 of file dataset.h. References dataset::d, EPSILON, and randu.

dataset* random_sample (  UINT  n  )  const [inline]

Generate a randomly sampled copy of the data set. Parameters: n  Number of random samples requested. Returns: A pointer to the new born data set. Samples are chosen with uniform probability. Definition at line 93 of file dataset.h. References dataset::d, and randu.

void replace (  UINT  i, const Tx &  _x, const Ty &  _y )  [inline]

Definition at line 163 of file dataset.h. Referenced by DataFeeder::linear_scale().

UINT size (   )  const [inline]

Definition at line 53 of file dataset.h. Referenced by LearnModel::exact_dimensions(), DataFeeder::linear_scale(), lemga::load_data(), DataFeeder::mean_var(), and DataFeeder::min_max().

const Tx& x (  UINT  i  )  const [inline]

Definition at line 56 of file dataset.h. Referenced by LearnModel::exact_dimensions(), DataFeeder::linear_scale(), DataFeeder::mean_var(), DataFeeder::min_max(), and LearnModel::set_dimensions().

const Ty& y (  UINT  i  )  const [inline]

Definition at line 57 of file dataset.h. Referenced by LearnModel::exact_dimensions(), DataFeeder::linear_scale(), and LearnModel::set_dimensions().

---

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

• dataset.h

---