IterativeFeatureSelector Class Reference

IterativeFeatureSelector - Apply bottom-up or top-down iteration for feature selection. More...

#include <FeatureProcess.h>

Inheritance diagram for IterativeFeatureSelector:

Public Member Functions
int	init (map< string, string > &mapper)
	The parsed fields from init command.
virtual void	init_defaults ()
void	init_bootstrap_cohort (MedBootstrapResult &bootstrapper, string &init)
void	init_bootstrap_params (MedBootstrapResult &bootstrapper, string &init)
virtual void	copy (FeatureProcessor *processor)
void	print_report (string &fileName)
void	retrace (MedFeatures &features, unordered_set< int > &ids, vector< string > &families_order, int start, int end)
void	retrace (MedFeatures &features, vector< string > &families_order, int start, int end)
ADD_SERIALIZATION_FUNCS(processor_type, predictor, predictor_params, predictor_params_vec, nfolds, folds, mode, rates_vec, cohort_params, bootstrap_params, msr_params, work_on_sets, required, ignored, numToSelect, selected, report, do_internal_cv, grouping_mode) private int	_learn (MedFeatures &features, unordered_set< int > &ids)
	Find set of selected features.
void	get_rates_vec ()
void	read_params_vec ()
void	get_features_families (MedFeatures &features, map< string, vector< string > > &featureFamilies)
void	prepare_for_iterations (MedBootstrapResult &bootstrapper, MedFeatures &features, vector< int > &folds, vector< vector< int > > &trainRows, vector< vector< int > > &testRows, vector< vector< float > > &trainLabels, vector< vector< MedSample > > &testSamples, MedFeatures &bootstrapFeatures)
void	pre_learn (MedFeatures &features, MedBootstrapResult &bootstrapper, map< string, vector< string > > &featureFamilies, vector< int > &orig_folds)
void	doTop2BottomSelection (MedFeatures &features, map< string, vector< string > > &featureFamilies, MedBootstrapResult &bootstrapper)
void	doBottom2TopSelection (MedFeatures &features, map< string, vector< string > > &featureFamilies, MedBootstrapResult &bootstrapper)
void	retraceTop2BottomSelection (MedFeatures &features, map< string, vector< string > > &featureFamilies, MedBootstrapResult &bootstrapper, vector< string > &order, int start, int end)
void	retraceBottom2TopSelection (MedFeatures &features, map< string, vector< string > > &featureFamilies, MedBootstrapResult &bootstrapper, vector< string > &order, int start, int end)
Public Member Functions inherited from FeatureSelector
virtual int	Learn (MedFeatures &features, unordered_set< int > &ids)
	Find set of selected features- Calls _learn function, and may be overrided directly.
virtual int	_apply (MedFeatures &features, unordered_set< int > &ids)
	Apply selection.
virtual int	_conditional_apply (MedFeatures &features, unordered_set< int > &ids, unordered_set< string > &out_req_features)
bool	is_selector ()
	allows testing if this feature processor is a selector
bool	are_features_affected (unordered_set< string > &out_req_features)
	check if a set of features is affected by the current processor
void	update_req_features_vec (unordered_set< string > &out_req_features, unordered_set< string > &in_req_features)
	update sets of required as input according to set required as output to processor
Public Member Functions inherited from FeatureProcessor
virtual string	select_learn_matrix (const vector< string > &matrix_tags) const
	Will be called before learn to create new version for the matrix if needed - in parallel of existing matrix.
virtual void	clear ()
void	init_defaults ()
virtual void	set_feature_name (const string &feature_name)
virtual string	get_feature_name ()
virtual void	get_feature_names (vector< string > &feature_names)
int	learn (MedFeatures &features)
	PostProcess of MedFeatures - on all ids.
int	learn (MedFeatures &features, unordered_set< int > &ids)
virtual int	_apply (MedFeatures &features, unordered_set< int > &ids, bool learning)
virtual int	_conditional_apply (MedFeatures &features, unordered_set< int > &ids, unordered_set< string > &req_features, bool learning)
int	apply (MedFeatures &features, bool learning)
	PostProcess of MedFeatures - on all or a subset of the ids calls virtaul function "_apply/_conditional_apply" for the specific implementation.
int	apply (MedFeatures &features, unordered_set< string > &req_features, bool learning)
int	apply (MedFeatures &features, unordered_set< int > &ids, bool learning)
int	apply (MedFeatures &features, unordered_set< int > &ids, unordered_set< string > &req_features, bool learning)
int	apply (MedFeatures &features)
int	apply (MedFeatures &features, unordered_set< string > &req_features)
int	apply (MedFeatures &features, unordered_set< int > &ids)
int	apply (MedFeatures &features, unordered_set< int > &ids, unordered_set< string > &req_features)
virtual int	init (void *processor_params)
virtual int	filter (unordered_set< string > &features)
	Filter according to a subset of features.
string	resolve_feature_name (MedFeatures &features, string substr)
	Utility : get corresponding name in MedFeatures.
void *	new_polymorphic (string derived_class_name)
	for polymorphic classes that want to be able to serialize/deserialize a pointer * to the derived class given its type one needs to implement this function to return a new to the derived class given its type (as in my_type)
size_t	get_processor_size ()
size_t	processor_serialize (unsigned char *blob)
virtual void	dprint (const string &pref, int rp_flag)
Public Member Functions inherited from SerializableObject
virtual int	version () const
	Relevant for serializations.
virtual string	my_class_name () const
	For better handling of serializations it is highly recommended that each SerializableObject inheriting class will implement the next method.
virtual void	serialized_fields_name (vector< string > &field_names) const
	The names of the serialized fields.
virtual void	pre_serialization ()
virtual void	post_deserialization ()
virtual size_t	get_size ()
	Gets bytes sizes for serializations.
virtual size_t	serialize (unsigned char *blob)
	Serialiazing object to blob memory. return number ob bytes wrote to memory.
virtual size_t	deserialize (unsigned char *blob)
	Deserialiazing blob to object. returns number of bytes read.
size_t	serialize_vec (vector< unsigned char > &blob)
size_t	deserialize_vec (vector< unsigned char > &blob)
virtual size_t	serialize (vector< unsigned char > &blob)
virtual size_t	deserialize (vector< unsigned char > &blob)
virtual int	read_from_file (const string &fname)
	read and deserialize model
virtual int	write_to_file (const string &fname)
	serialize model and write to file
virtual int	read_from_file_unsafe (const string &fname)
	read and deserialize model without checking version number - unsafe read
int	init_from_string (string init_string)
	Init from string.
int	init_params_from_file (string init_file)
int	init_param_from_file (string file_str, string &param)
int	update_from_string (const string &init_string)
virtual int	update (map< string, string > &map)
	Virtual to update object from parsed fields.
virtual string	object_json () const

Data Fields
string	predictor
	the predictor type - same as in the json file: qrf,lightgbm...
string	predictor_params
	the predictor parameters
string	predictor_params_file
	File with nFeatures-dependent predictor parameters.
int	nfolds = 5
	number of folds for cross-validation
bool	do_internal_cv = true
	use nfolds and create internal splits (if false, uses original samples' splits
vector< int >	folds
	if given, perform only subset of the possible 'nfolds' folds in cross-validation
string	mode = "top2bottom"
	'top2bottom' or 'bottom2top'
string	rates = "50:1,100:2,500:5,5000:10"
	instruction on rate of selection - comma separated pairs : #-bound:step
string	cohort_params
	cohort parameters for bootstrap performance evaluation (type:from,to/type:from,to/....)
string	bootstrap_params = "sample_per_pid:1"
	parameters for bootstrapping ('/' separaters)
string	msr_params = "AUC"
	measurements parameters for bootstrap performance evaluation
bool	work_on_sets = false
	work on sets of features according to signals
bool	group_to_sigs = false
	If true will group ungroupd_names to signals.
unordered_set< string >	ungroupd_names = { "Drug","RC","ICD9" }
	features-names (NAME in FTR_####.NAME) not to be grouped even in work_on_sets mode.
unordered_set< string >	ignored
	features to ignore in selection process
bool	verbose
	print all feature importance
string	progress_file_path = ""
	file path to progress file
string	grouping_mode = "BY_SIGNAL_CATEG"
	get also provide external file with the grouping
vector< int >	rates_vec
vector< string >	predictor_params_vec
string	measurement_name
vector< string >	report
Data Fields inherited from FeatureSelector
float	missing_value = (float)MED_MAT_MISSING_VALUE
	Missing Value.
unordered_set< string >	required
	Required Features.
vector< string >	selected
	Selected Features (ordered)
int	numToSelect = 0
	Target number to select (if 0, ignored)
int	numToSelectDelta = 0
	Delta around numToSelect. will search to find [numToSelect - numToSelectDelta, numToSelect + numToSelectDelta].
Data Fields inherited from FeatureProcessor
string	feature_name = "unset_feature_name"
	Feature name ( + name as appears in MedFeatures) ;.
string	resolved_feature_name
FeatureProcessorTypes	processor_type = FTR_PROCESS_LAST
int	learn_nthreads
int	clean_nthreads

Additional Inherited Members
Static Public Member Functions inherited from FeatureProcessor
static FeatureProcessor *	make_processor (string processor_name)
static FeatureProcessor *	make_processor (FeatureProcessorTypes type)
static FeatureProcessor *	make_processor (string processor_name, string params)
static FeatureProcessor *	make_processor (FeatureProcessorTypes type, string params)

Detailed Description

IterativeFeatureSelector - Apply bottom-up or top-down iteration for feature selection.

To Use this selector specify "iterative_selector" in the fp_type

Member Function Documentation

_learn()

int IterativeFeatureSelector::_learn ( MedFeatures &  features, unordered_set< int > &  ids  )

virtual

Find set of selected features.

Reimplemented from FeatureSelector.

copy()

virtual void IterativeFeatureSelector::copy ( FeatureProcessor *  processor )

inlinevirtual

Reimplemented from FeatureProcessor.

init()

int IterativeFeatureSelector::init ( map< string, string > &  mapper )

virtual

The parsed fields from init command.

        if (field == "missing_value") missing_value = med_stof(entry.second);
        else if (field == "predictor") predictor = entry.second;
        else if (field == "predictor_params") predictor_params = entry.second;
        else if (field == "importance_params") importance_params = entry.second;
        else if (field == "minStat") minStat = med_stof(entry.second);
        else if (field == "verbose") verbose = med_stoi(entry.second) > 0;
        else if (field == "numToSelect") numToSelect = med_stoi(entry.second);
        else if (field != "names" && field != "fp_type" && field != "tag")
            MLOG("Unknown parameter \'%s\' for ImportanceFeatureSelector\n", field.c_str());

[ImportanceFeatureSelector::init]

[ImportanceFeatureSelector::init]

Reimplemented from FeatureProcessor.

---

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

• Internal/MedProcessTools/MedProcessTools/FeatureProcess.h
• Internal/MedProcessTools/MedProcessTools/FeatureSelector.cpp
• Internal/MedProcessTools/MedProcessTools/IterativeSelector.cpp