class that specifies the machine learning algorithm to pair with Feat. More...

#include <ml.h>

Collaboration diagram for FT::Model::ML:

[legend]

Public Member Functions
	ML (string ml="LinearRidgeRegression", bool norm=true, bool classification=false, int n_classes=2)

void	init (bool assign_p_est=true)

	~ML ()

vector< float >	get_weights (bool norm_adjust=true) const

shared_ptr< CLabels >	fit (const MatrixXf &X, const VectorXf &y, const Parameters &params, bool &pass, const vector< char > &dtypes=vector< char >())

VectorXf	fit_vector (const MatrixXf &X, const VectorXf &y, const Parameters &params, bool &pass, const vector< char > &dtypes=vector< char >())

shared_ptr< CLabels >	predict (const MatrixXf &X, bool print=false)

VectorXf	predict_vector (const MatrixXf &X)

ArrayXXf	predict_proba (const MatrixXf &X)

VectorXf	labels_to_vector (const shared_ptr< CLabels > &labels)
	utility function to convert CLabels types to VectorXd types. More...

shared_ptr< CLabels >	retrieve_labels (CDenseFeatures< float64_t > *features, bool proba, bool &pass)
	returns labels of a fitted model estimating on features More...

void	set_dtypes (const vector< char > &dtypes)

float	get_bias (bool norm_adjust=true) const
	returns bias for linear machines More...

void	set_bias (float b)

shared_ptr< CLabels >	fit_tune (MatrixXf &X, VectorXf &y, const Parameters &params, bool &pass, const vector< char > &dtypes=vector< char >(), bool set_default=false)
	tune algorithm parameters More...

Public Attributes
std::map< string, ML_TYPE >	ml_hash

shared_ptr< sh::CMachine >	p_est
	pointer to the ML object More...

ML_TYPE	ml_type
	user specified ML type More...

string	ml_str
	user specified ML type (string) More...

sh::EProblemType	prob_type
	type of learning problem; binary, multiclass or regression More...

Normalizer	N
	normalization More...

int	max_train_time
	max seconds allowed for training More...

bool	normalize
	control whether ML normalizes its input before training More...

float	C

Private Attributes
vector< char >	dtypes

Detailed Description

class that specifies the machine learning algorithm to pair with Feat.

Definition at line 79 of file ml.h.

Constructor & Destructor Documentation

◆ ML()

FT::Model::ML::ML	(	string	ml = `"LinearRidgeRegression"`,
		bool	norm = `true`,
		bool	classification = `false`,
		int	n_classes = `2`
	)

use string to specify a desired ML algorithm from shogun.

Definition at line 21 of file ml.cc.

◆ ~ML()

FT::Model::ML::~ML ( )

Definition at line 187 of file ml.cc.

Member Function Documentation

◆ fit()

shared_ptr< CLabels > FT::Model::ML::fit	(	const MatrixXf &	X,
		const VectorXf &	y,
		const Parameters &	params,
		bool &	pass,
		const vector< char > &	dtypes = `vector<char>()`
	)

Trains ml on X, y to generate output yhat = f(X).

Parameters

	X	n_features x n_samples matrix
	y	n_samples vector of training labels
	params	feat parameters
[out]	pass	returns True if fit was successful, False if not
	dtypes	the data types of features in X

Returns: yhat: n_samples vector of outputs

Definition at line 282 of file ml.cc.

◆ fit_tune()

shared_ptr< CLabels > FT::Model::ML::fit_tune	(	MatrixXf &	X,
		VectorXf &	y,
		const Parameters &	params,
		bool &	pass,
		const vector< char > &	dtypes = `vector<char>()`,
		bool	set_default = `false`
	)

tune algorithm parameters

Definition at line 676 of file ml.cc.

◆ fit_vector()

VectorXf FT::Model::ML::fit_vector	(	const MatrixXf &	X,
		const VectorXf &	y,
		const Parameters &	params,
		bool &	pass,
		const vector< char > &	dtypes = `vector<char>()`
	)

Definition at line 413 of file ml.cc.

◆ get_bias()

float FT::Model::ML::get_bias ( bool norm_adjust = true ) const

returns bias for linear machines

Definition at line 565 of file ml.cc.

◆ get_weights()

vector< float > FT::Model::ML::get_weights ( bool norm_adjust = true ) const

Returns: weight vector from model.

Definition at line 211 of file ml.cc.

◆ init()

void FT::Model::ML::init ( bool assign_p_est = true )

Definition at line 71 of file ml.cc.

◆ labels_to_vector()

VectorXf FT::Model::ML::labels_to_vector ( const shared_ptr< CLabels > & labels )

utility function to convert CLabels types to VectorXd types.

Definition at line 539 of file ml.cc.

◆ predict()

shared_ptr< CLabels > FT::Model::ML::predict	(	const MatrixXf &	X,
		bool	print = `false`
	)

Definition at line 423 of file ml.cc.

◆ predict_proba()

ArrayXXf FT::Model::ML::predict_proba ( const MatrixXf & X )

Definition at line 499 of file ml.cc.

◆ predict_vector()

VectorXf FT::Model::ML::predict_vector ( const MatrixXf & X )

Definition at line 492 of file ml.cc.

◆ retrieve_labels()

shared_ptr< CLabels > FT::Model::ML::retrieve_labels	(	CDenseFeatures< float64_t > *	features,
		bool	proba,
		bool &	pass
	)

returns labels of a fitted model estimating on features

Definition at line 614 of file ml.cc.

◆ set_bias()

void FT::Model::ML::set_bias ( float b )

Definition at line 600 of file ml.cc.

◆ set_dtypes()

void FT::Model::ML::set_dtypes ( const vector< char > & dtypes )

Definition at line 192 of file ml.cc.

Member Data Documentation

◆ C

float FT::Model::ML::C

Definition at line 146 of file ml.h.

◆ dtypes

vector<char> FT::Model::ML::dtypes

private

Definition at line 149 of file ml.h.

◆ max_train_time

int FT::Model::ML::max_train_time

max seconds allowed for training

Definition at line 143 of file ml.h.

◆ ml_hash

std::map<string, ML_TYPE> FT::Model::ML::ml_hash

Definition at line 92 of file ml.h.

◆ ml_str

string FT::Model::ML::ml_str

user specified ML type (string)

Definition at line 139 of file ml.h.

◆ ml_type

ML_TYPE FT::Model::ML::ml_type

user specified ML type

Definition at line 138 of file ml.h.

◆ N

Normalizer FT::Model::ML::N

normalization

Definition at line 142 of file ml.h.

◆ normalize

bool FT::Model::ML::normalize

control whether ML normalizes its input before training

Definition at line 144 of file ml.h.

◆ p_est

shared_ptr<sh::CMachine> FT::Model::ML::p_est

pointer to the ML object

Definition at line 137 of file ml.h.

◆ prob_type

sh::EProblemType FT::Model::ML::prob_type

type of learning problem; binary, multiclass or regression

Definition at line 140 of file ml.h.

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

ml.h
ml.cc

Public Member Functions

Public Attributes

Private Attributes