DatasetFeature Class Reference

Metadata about a particular dataset feature (attribute). More...

Collaboration diagram for DatasetFeature:

Public Types
enum	Type { CONTINUOUS, INTEGER, NOMINAL, CONTINUOUS, INTEGER, NOMINAL }
	Feature types supported. More...
enum	Type { CONTINUOUS, INTEGER, NOMINAL, CONTINUOUS, INTEGER, NOMINAL }
	Feature types supported. More...
Public Member Functions
void	Format (double value, String *result) const
	Creates a text version of the value based on the type.
void	Format (double value, String *result) const
	Creates a text version of the value based on the type.
void	InitContinuous (const char *name_in)
	Initialize to be a continuous feature.
void	InitContinuous (const char *name_in)
	Initialize to be a continuous feature.
void	InitInteger (const char *name_in)
	Initializes to an integer type.
void	InitInteger (const char *name_in)
	Initializes to an integer type.
void	InitNominal (const char *name_in)
	Initializes to a nominal type.
void	InitNominal (const char *name_in)
	Initializes to a nominal type.
index_t	n_values () const
	The number of nominal values.
index_t	n_values () const
	The number of nominal values.
const String &	name () const
	Gets what the feature is named.
const String &	name () const
	Gets what the feature is named.
success_t	Parse (const char str, double d) const
	Parses a string into the particular value.
success_t	Parse (const char str, double d) const
	Parses a string into the particular value.
Type	type () const
	Identifies the type of feature.
Type	type () const
	Identifies the type of feature.
const String &	value_name (int value) const
	Returns the name of a particular nominal value, given its index.
const String &	value_name (int value) const
	Returns the name of a particular nominal value, given its index.
ArrayList< String > &	value_names ()
	Gets the array of value names.
ArrayList< String > &	value_names ()
	Gets the array of value names.

Detailed Description

Metadata about a particular dataset feature (attribute).

Supports nominal, continuous, and integer values.

Definition at line 59 of file dataset.h.

Member Enumeration Documentation

enum DatasetFeature::Type

Feature types supported.

Enumerator:

CONTINUOUS	Real-valued data.
INTEGER	Integer valued data.
NOMINAL	Discrete data, each of which has a "name".
CONTINUOUS	Real-valued data.
INTEGER	Integer valued data.
NOMINAL	Discrete data, each of which has a "name".

Definition at line 64 of file dataset.h.

enum DatasetFeature::Type

Feature types supported.

Enumerator:

CONTINUOUS	Real-valued data.
INTEGER	Integer valued data.
NOMINAL	Discrete data, each of which has a "name".
CONTINUOUS	Real-valued data.
INTEGER	Integer valued data.
NOMINAL	Discrete data, each of which has a "name".

Definition at line 64 of file dataset.h.

Member Function Documentation

void DatasetFeature::Format	(	double	value,
		String *	result
	)			const

Creates a text version of the value based on the type.

Continuous parameters are printed in floating point, and integers are shown as integers. For nominal, the value_name(int(value)) is shown. NaN (missing data) is always shown as '?'.

Parameters:

	value	the value to format
	result	this will be initialized to the formatted text

void DatasetFeature::Format	(	double	value,
		String *	result
	)			const

Creates a text version of the value based on the type.

Continuous parameters are printed in floating point, and integers are shown as integers. For nominal, the value_name(int(value)) is shown. NaN (missing data) is always shown as '?'.

Parameters:

	value	the value to format
	result	this will be initialized to the formatted text

Definition at line 49 of file dataset.cc.

References CONTINUOUS, String::Copy(), String::InitSprintf(), INTEGER, NOMINAL, and value_name().

Referenced by Dataset::Format().

void DatasetFeature::InitContinuous ( const char * name_in ) [inline]

Initialize to be a continuous feature.

Parameters:

name_in

the name of the feature

Definition at line 107 of file dataset.h.

References CONTINUOUS.

void DatasetFeature::InitContinuous ( const char * name_in ) [inline]

Initialize to be a continuous feature.

Parameters:

name_in

the name of the feature

Definition at line 107 of file dataset.h.

References CONTINUOUS.

void DatasetFeature::InitInteger ( const char * name_in ) [inline]

Initializes to an integer type.

Parameters:

name_in

the name of the feature

Definition at line 117 of file dataset.h.

References INTEGER.

void DatasetFeature::InitInteger ( const char * name_in ) [inline]

Initializes to an integer type.

Parameters:

name_in

the name of the feature

Definition at line 117 of file dataset.h.

References INTEGER.

void DatasetFeature::InitNominal ( const char * name_in ) [inline]

Initializes to a nominal type.

The value_names list starts empty, so you need to add the name of each feature to this. (The dataset reading functions will do this for you).

Parameters:

name_in

the name of the feature

Definition at line 131 of file dataset.h.

References NOMINAL.

void DatasetFeature::InitNominal ( const char * name_in ) [inline]

Initializes to a nominal type.

The value_names list starts empty, so you need to add the name of each feature to this. (The dataset reading functions will do this for you).

Parameters:

name_in

the name of the feature

Definition at line 131 of file dataset.h.

References NOMINAL.

Referenced by DatasetInfo::InitFromArff().

index_t DatasetFeature::n_values ( ) const [inline]

The number of nominal values.

The values 0 to n_values() - 1 are valid. This will return zero for CONTINUOUS and INTEGER types.

Returns:: the number of nominal values

Definition at line 200 of file dataset.h.

References ArrayList< TElem >::size().

index_t DatasetFeature::n_values ( ) const [inline]

The number of nominal values.

The values 0 to n_values() - 1 are valid. This will return zero for CONTINUOUS and INTEGER types.

Returns:: the number of nominal values

Definition at line 200 of file dataset.h.

References ArrayList< TElem >::size().

Referenced by SimpleCrossValidator< TClassifier >::Init(), and DatasetInfo::WriteArffHeader().

const String& DatasetFeature::name ( ) const [inline]

Gets what the feature is named.

Returns:: the name of the feature; for point, "Age" or "X Position"

Definition at line 167 of file dataset.h.

const String& DatasetFeature::name ( ) const [inline]

Gets what the feature is named.

Returns:: the name of the feature; for point, "Age" or "X Position"

Definition at line 167 of file dataset.h.

Referenced by DatasetInfo::WriteArffHeader().

success_t DatasetFeature::Parse	(	const char *	str,
		double *	d
	)			const

Parses a string into the particular value.

Integers and continuous are parsed using the normal functions. For nominal, the entry

If an invalid parse occurs, such as a mal-formatted number or a nominal value not in the list, SUCCESS_FAIL will be returned.

Parameters:

	str	the string to parse
	d	where to store the result

success_t DatasetFeature::Parse	(	const char *	str,
		double *	d
	)			const

Parses a string into the particular value.

Integers and continuous are parsed using the normal functions. For nominal, the entry

If an invalid parse occurs, such as a mal-formatted number or a nominal value not in the list, SUCCESS_FAIL will be returned.

Parameters:

	str	the string to parse
	d	where to store the result

Definition at line 72 of file dataset.cc.

References CONTINUOUS, DBL_NAN, INTEGER, NOMINAL, and ArrayList< TElem >::size().

Type DatasetFeature::type ( ) const [inline]

Identifies the type of feature.

Returns:: whether this is DatasetFeature::CONTINUOUS, INTEGER, or NOMINAL

Definition at line 176 of file dataset.h.

Type DatasetFeature::type ( ) const [inline]

Identifies the type of feature.

Returns:: whether this is DatasetFeature::CONTINUOUS, INTEGER, or NOMINAL

Definition at line 176 of file dataset.h.

Referenced by SimpleCrossValidator< TClassifier >::Init(), and DatasetInfo::WriteArffHeader().

const String& DatasetFeature::value_name ( int value ) const [inline]

Returns the name of a particular nominal value, given its index.

The first nominal value is 0, the second is 1, etc.

Parameters:

value

the number of the value

Definition at line 187 of file dataset.h.

References NOMINAL.

const String& DatasetFeature::value_name ( int value ) const [inline]

Returns the name of a particular nominal value, given its index.

The first nominal value is 0, the second is 1, etc.

Parameters:

value

the number of the value

Definition at line 187 of file dataset.h.

References NOMINAL.

Referenced by Format(), and DatasetInfo::WriteArffHeader().

ArrayList<String>& DatasetFeature::value_names ( ) [inline]

Gets the array of value names.

Useful for creating a nominal feature yourself.

Returns:: a mutable array of value names

Definition at line 211 of file dataset.h.

ArrayList<String>& DatasetFeature::value_names ( ) [inline]

Gets the array of value names.

Useful for creating a nominal feature yourself.

Returns:: a mutable array of value names

Definition at line 211 of file dataset.h.

Referenced by DatasetInfo::InitFromArff().

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

fastlib/data/dataset.h
include/fastlib/data/dataset.h
fastlib/data/dataset.cc

DatasetFeature Class Reference

Public Types

Public Member Functions

Detailed Description

Member Enumeration Documentation

Member Function Documentation