txeo::DataTableNorm< T > Class Template Reference

A normalizer for DataTable objects that handles feature scaling. More...

#include <DataTableNorm.h>

Collaboration diagram for txeo::DataTableNorm< T >:

Public Member Functions
	DataTableNorm ()=default
	DataTableNorm (const DataTableNorm &)=delete
	DataTableNorm (DataTableNorm &&)=default
DataTableNorm &	operator= (const DataTableNorm &)=delete
DataTableNorm &	operator= (DataTableNorm &&)=default
	~DataTableNorm ()=default
	DataTableNorm (const txeo::DataTable< T > &data, txeo::NormalizationType type=txeo::NormalizationType::MIN_MAX)
	Construct a new DataTableNorm object with associated DataTable.
const txeo::DataTable< T > &	data_table () const
	Get the associated DataTable.
void	set_data_table (const txeo::DataTable< T > &data)
	Set a new DataTable for normalization.
txeo::NormalizationType	type () const
	Get the current normalization type.
txeo::Matrix< T >	normalize (txeo::Matrix< T > &&x) const
	Normalize a matrix using rvalue semantics (move data)
txeo::Matrix< T >	normalize (const txeo::Matrix< T > &x) const
	Normalize a copy of the input matrix.
txeo::Matrix< T >	x_train_normalized ()
	Get normalized training data.
txeo::Matrix< T >	x_eval_normalized ()
	Get normalized evaluation data.
txeo::Matrix< T >	x_test_normalized ()
	Get normalized test data.

Detailed Description

template<typename T>
class txeo::DataTableNorm< T >

A normalizer for DataTable objects that handles feature scaling.

This class provides normalization functionality for machine learning datasets stored in DataTable format. It supports both Min-Max scaling and Z-Score standardization, computing normalization parameters from the training data.

Template Parameters

T	The data type of the table elements (e.g., float, double)

Note

The normalizer computes normalization parameters (min/max or mean/std) from the training portion of the DataTable. Make sure the DataTable is properly initialized with training data before use.

Definition at line 26 of file DataTableNorm.h.

Constructor & Destructor Documentation

DataTableNorm() [1/4]

template<typename T >

txeo::DataTableNorm< T >::DataTableNorm ( )

default

DataTableNorm() [2/4]

template<typename T >

txeo::DataTableNorm< T >::DataTableNorm ( const DataTableNorm< T > &  )

delete

DataTableNorm() [3/4]

template<typename T >

txeo::DataTableNorm< T >::DataTableNorm ( DataTableNorm< T > &&  )

default

~DataTableNorm()

template<typename T >

txeo::DataTableNorm< T >::~DataTableNorm ( )

default

DataTableNorm() [4/4]

template<typename T >

txeo::DataTableNorm< T >::DataTableNorm	(	const txeo::DataTable< T > &	data,
		txeo::NormalizationType	type = txeo::NormalizationType::MIN_MAX
	)

Construct a new DataTableNorm object with associated DataTable.

Parameters

data	Reference to the DataTable containing training data
type	Normalization type (MIN_MAX or Z_SCORE)

Exceptions

DataTableNormError

Example Usage:

txeo::DataTable<double> data = load_my_dataset();
txeo::DataTableNorm<double> normalizer(data, txeo::NormalizationType::Z_SCORE);
 
// Normalize a new sample
txeo::Matrix<double> sample = {{1.5}, {2.0}, {3.5}};
auto normalized_sample = normalizer.normalize(sample);

Member Function Documentation

data_table()

template<typename T >

const txeo::DataTable< T > & txeo::DataTableNorm< T >::data_table ( ) const

inline

Get the associated DataTable.

Returns

const txeo::DataTable<T>& Reference to the stored DataTable

Example Usage:

// Inspect the data table used for normalization
const auto& dt = normalizer.data_table();
std::cout << "Original dataset size: " << dt.x_train().rows() << std::endl;

Definition at line 68 of file DataTableNorm.h.

{ return *_data_table; }

normalize() [1/2]

template<typename T >

txeo::Matrix< T > txeo::DataTableNorm< T >::normalize ( const txeo::Matrix< T > &  x ) const

inline

Normalize a copy of the input matrix.

Parameters

x	Input matrix to normalize

Returns

txeo::Matrix<T> Normalized copy of the matrix

Example Usage:

txeo::Matrix<double> original = {{1.0}, {2.0}, {3.0}};
auto normalized = normalizer.normalize(original);  // Preserve original

Definition at line 127 of file DataTableNorm.h.

                                                          {
      return this->normalize(x.clone());
    };

Here is the call graph for this function:

normalize() [2/2]

template<typename T >

txeo::Matrix< T > txeo::DataTableNorm< T >::normalize

(

txeo::Matrix< T > &&

x

)

const

Normalize a matrix using rvalue semantics (move data)

Parameters

x	Input matrix to normalize (will be moved from)

Returns

txeo::Matrix<T> Normalized matrix

Exceptions

DataTableNormError

If normalization parameters not initialized

Example Usage:

txeo::Matrix<double> large_matrix = generate_large_data();
// Efficiently normalize without copying
auto normalized = normalizer.normalize(std::move(large_matrix));

Here is the caller graph for this function:

operator=() [1/2]

template<typename T >

DataTableNorm & txeo::DataTableNorm< T >::operator= ( const DataTableNorm< T > &  )

delete

operator=() [2/2]

template<typename T >

DataTableNorm & txeo::DataTableNorm< T >::operator= ( DataTableNorm< T > &&  )

default

set_data_table()

template<typename T >

void txeo::DataTableNorm< T >::set_data_table

(

const txeo::DataTable< T > &

data

)

Set a new DataTable for normalization.

Parameters

data	New DataTable to use for normalization parameters

Exceptions

DataTableNormError

If DataTable is invalid

Example Usage:

txeo::DataTable<double> new_data = load_updated_dataset();
normalizer.set_data_table(new_data);  // Update normalization parameters

type()

template<typename T >

txeo::NormalizationType txeo::DataTableNorm< T >::type ( ) const

inline

Get the current normalization type.

Returns

txeo::NormalizationType Active normalization method

Example Usage:

if(normalizer.type() == txeo::NormalizationType::MIN_MAX) {
    std::cout << "Using Min-Max normalization" << std::endl;
}

Definition at line 96 of file DataTableNorm.h.

{ return _type; }

x_eval_normalized()

template<typename T >

txeo::Matrix< T > txeo::DataTableNorm< T >::x_eval_normalized

(

)

Get normalized evaluation data.

Returns

txeo::Matrix<T> Normalized evaluation dataset

Example Usage:

auto x_eval_norm = normalizer.x_eval_normalized();
auto eval_score = model.evaluate(x_eval_norm, normalizer.data_table().y_eval());

x_test_normalized()

template<typename T >

txeo::Matrix< T > txeo::DataTableNorm< T >::x_test_normalized

(

)

Get normalized test data.

Returns

txeo::Matrix<T> Normalized test dataset

Example Usage:

auto x_test_norm = normalizer.x_test_normalized();
auto test_acc = model.test(x_test_norm, normalizer.data_table().y_test());

x_train_normalized()

template<typename T >

txeo::Matrix< T > txeo::DataTableNorm< T >::x_train_normalized

(

)

Get normalized training data.

Returns

txeo::Matrix<T> Normalized training dataset

Exceptions

DataTableNormError

If training data not available

Example Usage:

auto x_train_norm = normalizer.x_train_normalized();
model.train(x_train_norm, normalizer.data_table().y_train());

---

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

• include/txeo/DataTableNorm.h