#include <opennn/dataset/tabular_dataset.h>

Inheritance diagram for opennn::TabularDataset:

Classes
struct	TrainingTransform

Public Types
enum class	MissingValuesMethod { Unuse , Mean , Median , Interpolation }

Public Types inherited from opennn::Dataset
enum class	Codification { UTF8 , SHIFT_JIS }

enum class	Separator { Space , Tab , Comma , Semicolon }

enum class	StorageMode { Matrix , BinaryFile , GPUPersistantData }

Public Member Functions
	TabularDataset (const Index=0, const Shape &={0}, const Shape &={0})

	TabularDataset (const filesystem::path &, const string &, bool has_header=true, bool has_sample_ids=false, const Codification &=Codification::UTF8)

Index	get_samples_number () const noexcept override

MatrixR	get_data (const string &sample_role, const string &variable_role) const

MatrixR	get_data_from_indices (const vector< Index > &, const vector< Index > &) const

MatrixR	get_variable_data (Index) const

MatrixR	get_variable_data (Index, const vector< Index > &) const

MatrixR	get_feature_data (const string &) const

void	configure_forecasting (Index past, Index future=1, bool multi_target=false)

void	clear_forecasting ()

bool	is_forecasting () const noexcept

Index	get_past_time_steps () const noexcept

Index	get_future_time_steps () const noexcept

bool	get_multi_target () const noexcept

Tensor3	get_sequence_data (const string &sample_role, const string &feature_role) const

MatrixR	calculate_autocorrelations (Index=10) const

Tensor3	calculate_cross_correlations (Index=10) const

vector< Variable >	get_model_input_variables () const override

bool	sample_order_matters () const noexcept override

void	resize_input_shape (Index) override

void	set (Index=0, const Shape &={}, const Shape &={})

void	set (const filesystem::path &, const string &, bool new_has_header=true, bool new_has_ids=false, const Codification &=Codification::UTF8)

void	set (const filesystem::path &)

void	set_storage_mode (StorageMode) override

void	set_binary_cache_path (const filesystem::path &)

vector< string >	get_feature_scalers (const string &) const

void	set_variable_scalers (const string &)

void	set_variable_scalers (const vector< string > &)

void	set_default_variable_scalers ()

MissingValuesMethod	get_missing_values_method () const

string	get_missing_values_method_string () const

const string &	get_missing_values_label () const

Index	get_missing_values_number () const

const NumberFormat &	get_number_format () const

void	set_number_format (const NumberFormat &new_number_format)

void	set_number_format_auto ()

void	set_missing_values_label (string label)

void	set_missing_values_method (const MissingValuesMethod &method)

void	set_missing_values_method (const string &)

void	scrub_missing_values () override

void	calculate_missing_values_statistics ()

void	impute_missing_values_statistic (const MissingValuesMethod &)

void	reuse_input_incomplete_rows_binary ()

virtual void	impute_missing_values_unuse ()

virtual void	impute_missing_values_interpolate ()

vector< string >	unuse_uncorrelated_variables (const float=0.25f)

vector< string >	unuse_collinear_variables (const float=0.95f)

vector< string >	unuse_least_correlated_variables (const Index inputs_to_keep)

vector< Descriptives >	calculate_feature_descriptives () const

vector< Descriptives >	calculate_feature_descriptives (const string &) const

vector< Descriptives >	calculate_feature_descriptives (const string &, const vector< Index > &) const

vector< Descriptives >	calculate_variable_descriptives_positive_samples () const

vector< Descriptives >	calculate_variable_descriptives_negative_samples () const

vector< Descriptives >	calculate_variable_descriptives_categories (Index) const

vector< Histogram >	calculate_variable_distributions (const Index=10) const

vector< BoxPlot >	calculate_variables_box_plots () const

Tensor< Correlation, 2 >	calculate_input_variable_correlations (Correlation(*)(const MatrixR &, const MatrixR &), Correlation::Method, const string &) const

Tensor< Correlation, 2 >	calculate_input_variable_pearson_correlations () const

Tensor< Correlation, 2 >	calculate_input_variable_spearman_correlations () const

Tensor< Correlation, 2 >	calculate_input_target_variable_correlations (Correlation(*)(const MatrixR &, const MatrixR &), const string &) const

Tensor< Correlation, 2 >	calculate_input_target_variable_pearson_correlations () const

Tensor< Correlation, 2 >	calculate_input_target_variable_spearman_correlations () const

MatrixR	calculate_input_target_correlation_values () const override

FeatureScaling	calculate_used_feature_scaling (VariableRole) const override

vector< Descriptives >	scale_data ()

vector< Descriptives >	scale_features (const string &)

void	unscale_features (const string &, const vector< Descriptives > &)

FeatureScaling	prepare_training_scaling (VariableRole, const FeatureScaling &, Index) override

void	clear_training_scaling () noexcept override

void	enable_device_residency () override

VectorI	calculate_target_distribution () const override

vector< vector< Index > >	calculate_Tukey_outliers (float cleaning_parameter=1.5f, bool replace_with_nan=false)

vector< vector< Index > >	replace_Tukey_outliers_with_NaN (const float=1.5f)

bool	has_nan () const override

bool	has_nan_row (Index row_index) const

VectorI	count_nans_per_variable () const

Index	count_rows_with_nan () const

Index	count_nan () const

void	set_data_random ()

void	set_data_integer (const Index)

void	set_data_binary_classification ()

void	from_JSON (const JsonDocument &) override

void	to_JSON (JsonWriter &) const override

void	read_csv ()

void	fill_inputs (const vector< Index > &, const vector< Index > &, float *, FillMode, ColumnContiguity column_contiguity=ColumnContiguity::Unknown) const override

void	fill_decoder (const vector< Index > &, const vector< Index > &, float *, FillMode, ColumnContiguity column_contiguity=ColumnContiguity::Unknown) const override

void	fill_targets (const vector< Index > &, const vector< Index > &, float *, FillMode, ColumnContiguity column_contiguity=ColumnContiguity::Unknown) const override

void	fill_batch (Batch &, const vector< Index > &, const FeatureSelection &, FillMode) const override

Index	get_samples_number (SampleRole role_type) const

Index	get_samples_number (string_view role) const

const MatrixR &	get_data () const noexcept

virtual void	set_storage_mode (const string &)

Public Member Functions inherited from opennn::Dataset
virtual	~Dataset ()=default

Index	get_samples_number (SampleRole role_type) const

Index	get_samples_number (string_view role) const

Index	get_used_samples_number () const

vector< Index >	get_sample_indices (SampleRole) const

vector< Index >	get_sample_indices (string_view role) const

vector< Index >	get_used_sample_indices () const

const vector< SampleRole > &	get_sample_roles () const noexcept

void	set_fold_split (const vector< Index > &training, const vector< Index > &validation)

void	clear_fold_split () noexcept

Index	get_variables_number () const noexcept

Index	get_variables_number (VariableRole) const

Index	get_variables_number (string_view role) const

Index	get_used_variables_number () const

const vector< Variable > &	get_variables () const noexcept

vector< Variable >	get_variables (VariableRole) const

vector< Variable >	get_variables (string_view role) const

Index	get_variable_index (const string &) const

Index	get_variable_index (const Index) const

vector< Index >	get_variable_indices (VariableRole) const

vector< Index >	get_variable_indices (string_view role) const

vector< Index >	get_used_variables_indices () const

vector< string >	get_variable_names () const

vector< string >	get_variable_names (VariableRole) const

vector< string >	get_variable_names (string_view role) const

VariableType	get_variable_type (const Index index) const

Index	get_features_number () const

Index	get_features_number (VariableRole) const

Index	get_features_number (string_view role) const

vector< string >	get_feature_names () const

vector< string >	get_feature_names (VariableRole) const

vector< string >	get_feature_names (string_view role) const

vector< vector< Index > >	get_feature_indices () const

vector< Index >	get_feature_indices (const Index) const

vector< Index >	get_feature_indices (VariableRole) const

vector< Index >	get_feature_indices (string_view role) const

vector< Index >	get_used_feature_indices () const

Shape	get_shape (VariableRole) const

Shape	get_shape (string_view role) const

void	get_batches (const vector< Index > &, Index, Shuffle, vector< vector< Index > > &, optional< unsigned > shuffle_seed=nullopt) const

const vector< vector< string > > &	get_data_file_preview () const noexcept

const filesystem::path &	get_data_path () const noexcept

void	set_cache_directory (const filesystem::path &new_cache_directory)

StorageMode	get_storage_mode () const noexcept

string	get_storage_mode_string () const

string	get_separator_string () const

string	get_separator_name () const

string	get_codification_string () const

bool	get_display () const noexcept

bool	is_empty () const noexcept

Shape	get_input_shape () const noexcept

Shape	get_target_shape () const noexcept

void	record_memory (const string &stage) const

const MatrixR &	get_data () const noexcept

void	set_data (const MatrixR &)

void	set_data (MatrixR &&)

void	set_data_constant (float new_value)

void	disable_device_residency ()

bool	is_device_resident () const noexcept

bool	requests_device_residency () const noexcept

bool	uses_device_residency () const noexcept

const float *	get_device_data () const

Index	get_device_data_columns () const noexcept

virtual const TokenizerOperator *	get_training_tokenizer () const

virtual const TokenizerOperator *	get_training_tokenizer (VariableRole role) const

void	set_sample_roles (SampleRole)

void	set_sample_roles (string_view role)

void	set_sample_role (Index, SampleRole)

void	set_sample_role (Index index, string_view role)

void	set_sample_roles (const vector< string > &)

void	set_sample_roles (const vector< Index > &, SampleRole)

void	set_sample_roles (const vector< Index > &indices, string_view role)

void	set_default_variable_names ()

void	set_default_variable_roles ()

void	set_variable_roles (const vector< string > &)

void	set_variable_indices (const vector< Index > &, const vector< Index > &)

void	set_input_variables_unused ()

void	set_variable_role (Index, VariableRole)

void	set_variable_role (Index index, string_view role)

void	set_variable_role (const string &, VariableRole)

void	set_variable_role (const string &name, string_view role)

void	set_variable_names (const vector< string > &)

void	set_variables_number (const Index new_size)

void	set_variable_roles (VariableRole)

void	set_variable_roles (string_view role)

void	set_shape (VariableRole, const Shape &)

void	set_shape (string_view role, const Shape &shape)

virtual void	set_data_path (const filesystem::path &new_data_path)

void	set_has_header (bool new_has_header)

void	set_has_ids (bool new_has_ids)

bool	get_header_line () const

bool	get_has_sample_ids () const

const vector< string > &	get_sample_ids () const

VectorI	filter_data (const VectorR &, const VectorR &)

bool	has_categorical_variables () const

void	set_separator (const Separator &new_separator)

void	set_separator_string (const string &)

void	set_separator_name (const string &)

void	set_codification (const Codification &new_codification)

void	set_codification (const string &)

void	set_display (bool new_display)

bool	has_validation () const

void	split_samples (const float training_ratio=0.6f, float validation_ratio=0.2f, float testing_ratio=0.2f, bool shuffle=true)

void	split_samples_sequential (const float training_ratio=0.6f, float validation_ratio=0.2f, float testing_ratio=0.2f)

void	split_samples_random (const float training_ratio=0.6f, float validation_ratio=0.2f, float testing_ratio=0.2f)

pair< Index, Index >	count_binary_targets (const string &sample_role) const

void	save (const filesystem::path &) const

void	load (const filesystem::path &)

virtual bool	supports_bf16_inputs () const

FeatureSelection	get_feature_selection () const

bool	can_device_gather (const Batch &, const FeatureSelection &) const

Protected Member Functions
void	set_forecasting_window (Index, Index, bool)

void	refresh_forecasting_roles ()

pair< vector< Index >, Index >	correlation_lags (Index) const

void	fill_window_features (const vector< Index > &, const vector< Index > &, float *, Index row_offset, Index steps, bool feature_major, ColumnContiguity) const

void	missing_values_to_JSON (JsonWriter &) const

void	missing_values_from_JSON (const Json *) override

void	resize_data_from_JSON (Index) override

filesystem::path	cache_file_path () const

void	fill_features (const vector< Index > &, const vector< Index > &, float *, ColumnContiguity column_contiguity=ColumnContiguity::Unknown) const

void	fill_from_binary_cache (const vector< Index > &, const vector< Index > &, float *, ColumnContiguity column_contiguity=ColumnContiguity::Unknown) const

float	apply_training_scaling (Index, float) const

void	apply_training_scaling (const vector< Index > &, float *, Index) const

vector< Index >	filter_used_samples_by_column (Index column, bool positive) const

vector< Descriptives >	calculate_variable_descriptives_samples (bool positive) const

vector< ScalerMethod >	get_feature_scaler_methods () const

vector< ScalerMethod >	get_feature_scaler_methods (VariableRole) const

void	unuse_samples_with_missing_targets (const vector< Index > &, const vector< Index > &)

void	on_used_samples_changed () override

void	clear_cache_derived_state ()

void	refresh_cache_statistics ()

vector< Descriptives >	compute_descriptives_streaming (const vector< Index > &) const

DateFormat	configure_csv_columns (vector< string_view > &, char, bool)

void	load_csv_data (const vector< string_view > &, char, bool, DateFormat)

DateFormat	infer_column_types (const vector< string_view > &, char, bool has_quotes=false)

void	apply_scaler (Index feature_index, ScalerMethod scaler, const Descriptives &descriptives, bool unscale)

Protected Member Functions inherited from opennn::Dataset
	Dataset ()=default

void	fill_batch_host (Batch &, const vector< Index > &sample_indices, const FeatureSelection &, FillMode) const

DeviceGather &	start_device_gather (Batch &, const vector< Index > &sample_indices, const FeatureSelection &) const

void	read_data_file_preview (const vector< string_view > &, char, bool has_quotes=false)

void	check_separators (string_view) const

void	samples_from_JSON (const Json *)

void	require_in_memory_data (string_view what) const

void	upload_device_matrix (const MatrixR &)

void	upload_device_samples ()

void	invalidate_data ()

const vector< SampleRole > &	active_sample_roles () const noexcept

void	variables_to_JSON (JsonWriter &) const

void	samples_to_JSON (JsonWriter &) const

void	preview_data_to_JSON (JsonWriter &) const

void	variables_from_JSON (const Json *)

void	preview_data_from_JSON (const Json *)

void	write_json_header (JsonWriter &, initializer_list< pair< const char *, Json > >) const

void	write_json_footer (JsonWriter &) const

void	read_json_blocks (const Json *)

Protected Attributes
Index	past_time_steps = 0

Index	future_time_steps = 1

bool	multi_target = false

string	missing_values_label = "NA"

NumberFormat	number_format

bool	number_format_automatic = true

MissingValuesMethod	missing_values_method = MissingValuesMethod::Mean

Index	missing_values_number = 0

VectorI	variables_missing_values_number

Index	rows_missing_values_number = 0

filesystem::path	cache_path

filesystem::path	cache_path_override

FileReader	cache_reader

Index	cache_columns_number = 0

vector< Descriptives >	cache_feature_descriptives

vector< Descriptives >	cache_transform_descriptives

vector< float >	cache_feature_replacement

vector< ScalerMethod >	cache_feature_transforms

vector< TrainingTransform >	training_transforms

Protected Attributes inherited from opennn::Dataset
StorageMode	storage_mode = StorageMode::Matrix

MatrixR	data

Buffer	data_device {Device::CUDA}

Index	device_data_columns = 0

Shape	input_shape

Shape	target_shape

Shape	decoder_shape

vector< SampleRole >	sample_roles

vector< string >	sample_ids

optional< vector< SampleRole > >	fold_split_roles

vector< Variable >	variables

filesystem::path	data_path

filesystem::path	cache_directory

Separator	separator = Separator::Comma

bool	has_header = false

bool	has_sample_ids = false

Codification	codification = Codification::UTF8

vector< vector< string > >	data_file_preview

bool	display = true

Member Enumeration Documentation

◆ MissingValuesMethod

enum class opennn::TabularDataset::MissingValuesMethod

strong

Enumerator
Unuse
Mean
Median
Interpolation

Constructor & Destructor Documentation

◆ TabularDataset() [1/2]

opennn::TabularDataset::TabularDataset	(	const Index	= 0,
		const Shape &	= {0},
		const Shape &	= {0} )

◆ TabularDataset() [2/2]

opennn::TabularDataset::TabularDataset	(	const filesystem::path &	,
		const string &	,
		bool	has_header = true,
		bool	has_sample_ids = false,
		const Codification &	= Codification::UTF8 )

Member Function Documentation

◆ apply_scaler()

void opennn::TabularDataset::apply_scaler	(	Index	feature_index,
		ScalerMethod	scaler,
		const Descriptives &	descriptives,
		bool	unscale )

protected

◆ apply_training_scaling() [1/2]

void opennn::TabularDataset::apply_training_scaling	(	const vector< Index > &	,
		float *	,
		Index	) const

protected

◆ apply_training_scaling() [2/2]

float opennn::TabularDataset::apply_training_scaling	(	Index	,
		float	) const

protected

◆ cache_file_path()

filesystem::path opennn::TabularDataset::cache_file_path ( ) const

protected

◆ calculate_autocorrelations()

MatrixR opennn::TabularDataset::calculate_autocorrelations ( Index = 10 ) const

◆ calculate_cross_correlations()

Tensor3 opennn::TabularDataset::calculate_cross_correlations ( Index = 10 ) const

◆ calculate_feature_descriptives() [1/3]

vector< Descriptives > opennn::TabularDataset::calculate_feature_descriptives ( ) const

◆ calculate_feature_descriptives() [2/3]

vector< Descriptives > opennn::TabularDataset::calculate_feature_descriptives ( const string & ) const

◆ calculate_feature_descriptives() [3/3]

vector< Descriptives > opennn::TabularDataset::calculate_feature_descriptives	(	const string &	,
		const vector< Index > &	) const

◆ calculate_input_target_correlation_values()

MatrixR opennn::TabularDataset::calculate_input_target_correlation_values ( ) const

inlineoverridevirtual

Reimplemented from opennn::Dataset.

◆ calculate_input_target_variable_correlations()

Tensor< Correlation, 2 > opennn::TabularDataset::calculate_input_target_variable_correlations	(	Correlation(*	)(const MatrixR &, const MatrixR &),
		const string &	) const

◆ calculate_input_target_variable_pearson_correlations()

Tensor< Correlation, 2 > opennn::TabularDataset::calculate_input_target_variable_pearson_correlations ( ) const

inline

◆ calculate_input_target_variable_spearman_correlations()

Tensor< Correlation, 2 > opennn::TabularDataset::calculate_input_target_variable_spearman_correlations ( ) const

inline

◆ calculate_input_variable_correlations()

Tensor< Correlation, 2 > opennn::TabularDataset::calculate_input_variable_correlations	(	Correlation(*	)(const MatrixR &, const MatrixR &),
		Correlation::Method	,
		const string &	) const

◆ calculate_input_variable_pearson_correlations()

Tensor< Correlation, 2 > opennn::TabularDataset::calculate_input_variable_pearson_correlations ( ) const

inline

◆ calculate_input_variable_spearman_correlations()

Tensor< Correlation, 2 > opennn::TabularDataset::calculate_input_variable_spearman_correlations ( ) const

inline

◆ calculate_missing_values_statistics()

void opennn::TabularDataset::calculate_missing_values_statistics ( )

◆ calculate_target_distribution()

VectorI opennn::TabularDataset::calculate_target_distribution ( ) const

overridevirtual

Reimplemented from opennn::Dataset.

◆ calculate_Tukey_outliers()

vector< vector< Index > > opennn::TabularDataset::calculate_Tukey_outliers	(	float	cleaning_parameter = 1.5f,
		bool	replace_with_nan = false )

◆ calculate_used_feature_scaling()

FeatureScaling opennn::TabularDataset::calculate_used_feature_scaling ( VariableRole ) const

overridevirtual

Reimplemented from opennn::Dataset.

◆ calculate_variable_descriptives_categories()

vector< Descriptives > opennn::TabularDataset::calculate_variable_descriptives_categories ( Index ) const

◆ calculate_variable_descriptives_negative_samples()

vector< Descriptives > opennn::TabularDataset::calculate_variable_descriptives_negative_samples ( ) const

inline

◆ calculate_variable_descriptives_positive_samples()

vector< Descriptives > opennn::TabularDataset::calculate_variable_descriptives_positive_samples ( ) const

inline

◆ calculate_variable_descriptives_samples()

vector< Descriptives > opennn::TabularDataset::calculate_variable_descriptives_samples ( bool positive ) const

protected

◆ calculate_variable_distributions()

vector< Histogram > opennn::TabularDataset::calculate_variable_distributions ( const Index = 10 ) const

◆ calculate_variables_box_plots()

vector< BoxPlot > opennn::TabularDataset::calculate_variables_box_plots ( ) const

◆ clear_cache_derived_state()

void opennn::TabularDataset::clear_cache_derived_state ( )

protected

◆ clear_forecasting()

void opennn::TabularDataset::clear_forecasting ( )

◆ clear_training_scaling()

void opennn::TabularDataset::clear_training_scaling ( )

overridevirtualnoexcept

Reimplemented from opennn::Dataset.

◆ compute_descriptives_streaming()

vector< Descriptives > opennn::TabularDataset::compute_descriptives_streaming ( const vector< Index > & ) const

protected

◆ configure_csv_columns()

DateFormat opennn::TabularDataset::configure_csv_columns	(	vector< string_view > &	,
		char	,
		bool	)

protected

◆ configure_forecasting()

void opennn::TabularDataset::configure_forecasting	(	Index	past,
		Index	future = 1,
		bool	multi_target = false )

◆ correlation_lags()

pair< vector< Index >, Index > opennn::TabularDataset::correlation_lags ( Index ) const

protected

◆ count_nan()

Index opennn::TabularDataset::count_nan ( ) const

inline

◆ count_nans_per_variable()

VectorI opennn::TabularDataset::count_nans_per_variable ( ) const

◆ count_rows_with_nan()

Index opennn::TabularDataset::count_rows_with_nan ( ) const

inline

◆ enable_device_residency()

void opennn::TabularDataset::enable_device_residency ( )

overridevirtual

Reimplemented from opennn::Dataset.

◆ fill_batch()

void opennn::TabularDataset::fill_batch	(	Batch &	,
		const vector< Index > &	,
		const FeatureSelection &	,
		FillMode	) const

overridevirtual

Reimplemented from opennn::Dataset.

◆ fill_decoder()

void opennn::TabularDataset::fill_decoder	(	const vector< Index > &	,
		const vector< Index > &	,
		float *	,
		FillMode	,
		ColumnContiguity	column_contiguity = ColumnContiguity::Unknown ) const

overridevirtual

Reimplemented from opennn::Dataset.

◆ fill_features()

void opennn::TabularDataset::fill_features	(	const vector< Index > &	,
		const vector< Index > &	,
		float *	,
		ColumnContiguity	column_contiguity = ColumnContiguity::Unknown ) const

protected

◆ fill_from_binary_cache()

void opennn::TabularDataset::fill_from_binary_cache	(	const vector< Index > &	,
		const vector< Index > &	,
		float *	,
		ColumnContiguity	column_contiguity = ColumnContiguity::Unknown ) const

protected

◆ fill_inputs()

void opennn::TabularDataset::fill_inputs	(	const vector< Index > &	,
		const vector< Index > &	,
		float *	,
		FillMode	,
		ColumnContiguity	column_contiguity = ColumnContiguity::Unknown ) const

overridevirtual

Reimplemented from opennn::Dataset.

◆ fill_targets()

void opennn::TabularDataset::fill_targets	(	const vector< Index > &	,
		const vector< Index > &	,
		float *	,
		FillMode	,
		ColumnContiguity	column_contiguity = ColumnContiguity::Unknown ) const

overridevirtual

Reimplemented from opennn::Dataset.

◆ fill_window_features()

void opennn::TabularDataset::fill_window_features	(	const vector< Index > &	,
		const vector< Index > &	,
		float *	,
		Index	row_offset,
		Index	steps,
		bool	feature_major,
		ColumnContiguity	) const

protected

◆ filter_used_samples_by_column()

vector< Index > opennn::TabularDataset::filter_used_samples_by_column	(	Index	column,
		bool	positive ) const

protected

◆ from_JSON()

void opennn::TabularDataset::from_JSON ( const JsonDocument & )

overridevirtual

Implements opennn::Dataset.

◆ get_data() [1/2]

const MatrixR & opennn::Dataset::get_data ( ) const

inlinenoexcept

◆ get_data() [2/2]

MatrixR opennn::TabularDataset::get_data	(	const string &	sample_role,
		const string &	variable_role ) const

inline

◆ get_data_from_indices()

MatrixR opennn::TabularDataset::get_data_from_indices	(	const vector< Index > &	,
		const vector< Index > &	) const

◆ get_feature_data()

MatrixR opennn::TabularDataset::get_feature_data ( const string & ) const

◆ get_feature_scaler_methods() [1/2]

vector< ScalerMethod > opennn::TabularDataset::get_feature_scaler_methods ( ) const

protected

◆ get_feature_scaler_methods() [2/2]

vector< ScalerMethod > opennn::TabularDataset::get_feature_scaler_methods ( VariableRole ) const

protected

◆ get_feature_scalers()

vector< string > opennn::TabularDataset::get_feature_scalers ( const string & ) const

◆ get_future_time_steps()

Index opennn::TabularDataset::get_future_time_steps ( ) const

inlinenoexcept

◆ get_missing_values_label()

const string & opennn::TabularDataset::get_missing_values_label ( ) const

inline

◆ get_missing_values_method()

MissingValuesMethod opennn::TabularDataset::get_missing_values_method ( ) const

inline

◆ get_missing_values_method_string()

string opennn::TabularDataset::get_missing_values_method_string ( ) const

◆ get_missing_values_number()

Index opennn::TabularDataset::get_missing_values_number ( ) const

inline

◆ get_model_input_variables()

vector< Variable > opennn::TabularDataset::get_model_input_variables ( ) const

overridevirtual

Reimplemented from opennn::Dataset.

◆ get_multi_target()

bool opennn::TabularDataset::get_multi_target ( ) const

inlinenoexcept

◆ get_number_format()

const NumberFormat & opennn::TabularDataset::get_number_format ( ) const

inline

◆ get_past_time_steps()

Index opennn::TabularDataset::get_past_time_steps ( ) const

inlinenoexcept

◆ get_samples_number() [1/3]

Index opennn::TabularDataset::get_samples_number ( ) const

inlineoverridevirtualnoexcept

Reimplemented from opennn::Dataset.

◆ get_samples_number() [2/3]

Index opennn::Dataset::get_samples_number ( SampleRole role_type ) const

inline

◆ get_samples_number() [3/3]

Index opennn::Dataset::get_samples_number ( string_view role ) const

inline

◆ get_sequence_data()

Tensor3 opennn::TabularDataset::get_sequence_data	(	const string &	sample_role,
		const string &	feature_role ) const

◆ get_variable_data() [1/2]

MatrixR opennn::TabularDataset::get_variable_data ( Index ) const

◆ get_variable_data() [2/2]

MatrixR opennn::TabularDataset::get_variable_data	(	Index	,
		const vector< Index > &	) const

◆ has_nan()

bool opennn::TabularDataset::has_nan ( ) const

overridevirtual

Reimplemented from opennn::Dataset.

◆ has_nan_row()

bool opennn::TabularDataset::has_nan_row ( Index row_index ) const

inline

◆ impute_missing_values_interpolate()

virtual void opennn::TabularDataset::impute_missing_values_interpolate ( )

virtual

◆ impute_missing_values_statistic()

void opennn::TabularDataset::impute_missing_values_statistic ( const MissingValuesMethod & )

◆ impute_missing_values_unuse()

virtual void opennn::TabularDataset::impute_missing_values_unuse ( )

virtual

◆ infer_column_types()

DateFormat opennn::TabularDataset::infer_column_types	(	const vector< string_view > &	,
		char	,
		bool	has_quotes = false )

protected

◆ is_forecasting()

bool opennn::TabularDataset::is_forecasting ( ) const

inlinenoexcept

◆ load_csv_data()

void opennn::TabularDataset::load_csv_data	(	const vector< string_view > &	,
		char	,
		bool	,
		DateFormat	)

protected

◆ missing_values_from_JSON()

void opennn::TabularDataset::missing_values_from_JSON ( const Json * )

overrideprotectedvirtual

Reimplemented from opennn::Dataset.

◆ missing_values_to_JSON()

void opennn::TabularDataset::missing_values_to_JSON ( JsonWriter & ) const

protected

◆ on_used_samples_changed()

void opennn::TabularDataset::on_used_samples_changed ( )

overrideprotectedvirtual

Reimplemented from opennn::Dataset.

◆ prepare_training_scaling()

FeatureScaling opennn::TabularDataset::prepare_training_scaling	(	VariableRole	,
		const FeatureScaling &	,
		Index	)

overridevirtual

Reimplemented from opennn::Dataset.

◆ read_csv()

void opennn::TabularDataset::read_csv ( )

◆ refresh_cache_statistics()

void opennn::TabularDataset::refresh_cache_statistics ( )

protected

◆ refresh_forecasting_roles()

void opennn::TabularDataset::refresh_forecasting_roles ( )

protected

◆ replace_Tukey_outliers_with_NaN()

vector< vector< Index > > opennn::TabularDataset::replace_Tukey_outliers_with_NaN ( const float = 1.5f )

◆ resize_data_from_JSON()

void opennn::TabularDataset::resize_data_from_JSON ( Index )

overrideprotectedvirtual

Reimplemented from opennn::Dataset.

◆ resize_input_shape()

void opennn::TabularDataset::resize_input_shape ( Index )

overridevirtual

Reimplemented from opennn::Dataset.

◆ reuse_input_incomplete_rows_binary()

void opennn::TabularDataset::reuse_input_incomplete_rows_binary ( )

◆ sample_order_matters()

bool opennn::TabularDataset::sample_order_matters ( ) const

inlineoverridevirtualnoexcept

Reimplemented from opennn::Dataset.

◆ scale_data()

vector< Descriptives > opennn::TabularDataset::scale_data ( )

◆ scale_features()

vector< Descriptives > opennn::TabularDataset::scale_features ( const string & )

◆ scrub_missing_values()

void opennn::TabularDataset::scrub_missing_values ( )

overridevirtual

Reimplemented from opennn::Dataset.

◆ set() [1/3]

void opennn::TabularDataset::set ( const filesystem::path & )

◆ set() [2/3]

void opennn::TabularDataset::set	(	const filesystem::path &	,
		const string &	,
		bool	new_has_header = true,
		bool	new_has_ids = false,
		const Codification &	= Codification::UTF8 )

◆ set() [3/3]

void opennn::TabularDataset::set	(	Index	= 0,
		const Shape &	= {},
		const Shape &	= {} )

◆ set_binary_cache_path()

void opennn::TabularDataset::set_binary_cache_path ( const filesystem::path & )

◆ set_data_binary_classification()

void opennn::TabularDataset::set_data_binary_classification ( )

◆ set_data_integer()

void opennn::TabularDataset::set_data_integer ( const Index )

◆ set_data_random()

void opennn::TabularDataset::set_data_random ( )

◆ set_default_variable_scalers()

void opennn::TabularDataset::set_default_variable_scalers ( )

◆ set_forecasting_window()

void opennn::TabularDataset::set_forecasting_window	(	Index	,
		Index	,
		bool	)

protected

◆ set_missing_values_label()

void opennn::TabularDataset::set_missing_values_label ( string label )

inline

◆ set_missing_values_method() [1/2]

void opennn::TabularDataset::set_missing_values_method ( const MissingValuesMethod & method )

inline

◆ set_missing_values_method() [2/2]

void opennn::TabularDataset::set_missing_values_method ( const string & )

◆ set_number_format()

void opennn::TabularDataset::set_number_format ( const NumberFormat & new_number_format )

inline

◆ set_number_format_auto()

void opennn::TabularDataset::set_number_format_auto ( )

inline

◆ set_storage_mode() [1/2]

virtual void opennn::Dataset::set_storage_mode ( const string & )

virtual

Reimplemented from opennn::Dataset.

◆ set_storage_mode() [2/2]

void opennn::TabularDataset::set_storage_mode ( StorageMode )

overridevirtual

Reimplemented from opennn::Dataset.

◆ set_variable_scalers() [1/2]

void opennn::TabularDataset::set_variable_scalers ( const string & )

◆ set_variable_scalers() [2/2]

void opennn::TabularDataset::set_variable_scalers ( const vector< string > & )

◆ to_JSON()

void opennn::TabularDataset::to_JSON ( JsonWriter & ) const

overridevirtual

Reimplemented from opennn::Dataset.

◆ unscale_features()

void opennn::TabularDataset::unscale_features	(	const string &	,
		const vector< Descriptives > &	)

◆ unuse_collinear_variables()

vector< string > opennn::TabularDataset::unuse_collinear_variables ( const float = 0.95f )

◆ unuse_least_correlated_variables()

vector< string > opennn::TabularDataset::unuse_least_correlated_variables ( const Index inputs_to_keep )

◆ unuse_samples_with_missing_targets()

void opennn::TabularDataset::unuse_samples_with_missing_targets	(	const vector< Index > &	,
		const vector< Index > &	)

protected

◆ unuse_uncorrelated_variables()

vector< string > opennn::TabularDataset::unuse_uncorrelated_variables ( const float = 0.25f )

Member Data Documentation

◆ cache_columns_number

Index opennn::TabularDataset::cache_columns_number = 0

protected

◆ cache_feature_descriptives

vector<Descriptives> opennn::TabularDataset::cache_feature_descriptives

protected

◆ cache_feature_replacement

vector<float> opennn::TabularDataset::cache_feature_replacement

protected

◆ cache_feature_transforms

vector<ScalerMethod> opennn::TabularDataset::cache_feature_transforms

protected

◆ cache_path

filesystem::path opennn::TabularDataset::cache_path

protected

◆ cache_path_override

filesystem::path opennn::TabularDataset::cache_path_override

protected

◆ cache_reader

FileReader opennn::TabularDataset::cache_reader

mutableprotected

◆ cache_transform_descriptives

vector<Descriptives> opennn::TabularDataset::cache_transform_descriptives

protected

◆ future_time_steps

Index opennn::TabularDataset::future_time_steps = 1

protected

◆ missing_values_label

string opennn::TabularDataset::missing_values_label = "NA"

protected

◆ missing_values_method

MissingValuesMethod opennn::TabularDataset::missing_values_method = MissingValuesMethod::Mean

protected

◆ missing_values_number

Index opennn::TabularDataset::missing_values_number = 0

protected

◆ multi_target

bool opennn::TabularDataset::multi_target = false

protected

◆ number_format

NumberFormat opennn::TabularDataset::number_format

protected

◆ number_format_automatic

bool opennn::TabularDataset::number_format_automatic = true

protected

◆ past_time_steps

Index opennn::TabularDataset::past_time_steps = 0

protected

◆ rows_missing_values_number

Index opennn::TabularDataset::rows_missing_values_number = 0

protected

◆ training_transforms

vector<TrainingTransform> opennn::TabularDataset::training_transforms

protected

◆ variables_missing_values_number

VectorI opennn::TabularDataset::variables_missing_values_number

protected

Classes

Public Types

Public Member Functions

Protected Member Functions

Protected Attributes

Member Enumeration Documentation

◆ MissingValuesMethod

Constructor & Destructor Documentation

◆ TabularDataset() [1/2]

◆ TabularDataset() [2/2]

Member Function Documentation

◆ apply_scaler()

◆ apply_training_scaling() [1/2]

◆ apply_training_scaling() [2/2]

◆ cache_file_path()

◆ calculate_autocorrelations()

◆ calculate_cross_correlations()

◆ calculate_feature_descriptives() [1/3]

◆ calculate_feature_descriptives() [2/3]

◆ calculate_feature_descriptives() [3/3]

◆ calculate_input_target_correlation_values()

◆ calculate_input_target_variable_correlations()

◆ calculate_input_target_variable_pearson_correlations()

◆ calculate_input_target_variable_spearman_correlations()

◆ calculate_input_variable_correlations()

◆ calculate_input_variable_pearson_correlations()

◆ calculate_input_variable_spearman_correlations()

◆ calculate_missing_values_statistics()

◆ calculate_target_distribution()

◆ calculate_Tukey_outliers()

◆ calculate_used_feature_scaling()

◆ calculate_variable_descriptives_categories()

◆ calculate_variable_descriptives_negative_samples()

◆ calculate_variable_descriptives_positive_samples()

◆ calculate_variable_descriptives_samples()

◆ calculate_variable_distributions()

◆ calculate_variables_box_plots()

◆ clear_cache_derived_state()

◆ clear_forecasting()

◆ clear_training_scaling()

◆ compute_descriptives_streaming()

◆ configure_csv_columns()

◆ configure_forecasting()

◆ correlation_lags()

◆ count_nan()

◆ count_nans_per_variable()

◆ count_rows_with_nan()

◆ enable_device_residency()

◆ fill_batch()

◆ fill_decoder()

◆ fill_features()

◆ fill_from_binary_cache()

◆ fill_inputs()

◆ fill_targets()

◆ fill_window_features()

◆ filter_used_samples_by_column()

◆ from_JSON()

◆ get_data() [1/2]

◆ get_data() [2/2]

◆ get_data_from_indices()

◆ get_feature_data()

◆ get_feature_scaler_methods() [1/2]

◆ get_feature_scaler_methods() [2/2]

◆ get_feature_scalers()

◆ get_future_time_steps()

◆ get_missing_values_label()

◆ get_missing_values_method()

◆ get_missing_values_method_string()

◆ get_missing_values_number()

◆ get_model_input_variables()

◆ get_multi_target()

◆ get_number_format()

◆ get_past_time_steps()

◆ get_samples_number() [1/3]

◆ get_samples_number() [2/3]

◆ get_samples_number() [3/3]

◆ get_sequence_data()

◆ get_variable_data() [1/2]

◆ get_variable_data() [2/2]

◆ has_nan()