File List

Here is a list of all documented files with brief descriptions:

_src_criteria
classifier.hpp	Defines interface for classifier implementations
classifier_extern.hpp	Defines Classifier_Extern — bridge delegating train/test to a user-supplied external command (persistent process, line protocol)
classifier_knn.hpp	Implements k-Nearest Neighbor classifier
classifier_lda.hpp	Defines Classifier_LDA — shared-covariance Gaussian (linear) discriminant for wrapper-based feature selection
classifier_linear_svm.hpp	Defines Classifier_Linear_SVM — dependency-free linear SVM (dual coordinate descent) for wrapper-based feature selection
classifier_logreg.hpp	Implements multinomial logistic regression classifier (dependency-free, deterministic; regularized damped-Newton training on internally standardized data)
classifier_multinom_naivebayes.hpp	Implements Naive-like Bayes classifier based on multinomial model
classifier_normal_bayes.hpp	Implements Bayes classifier based on normal (gaussian) model
classifier_rbf_svm.hpp	Defines Classifier_RBF_SVM — dependency-free RBF kernel SVM (kernel dual coordinate descent) for wrapper-based feature selection
classifier_rf.hpp	Defines Classifier_Random_Forest — dependency-free, deterministic-by-design Random Forest for wrapper-based feature selection
classifier_svm.hpp	Wraps external Support Vector Machine implementation (in LibSVM) to serve as FST4 classifier
criterion.hpp	Defines interface for feature selection criterion classes
criterion_extern_value.hpp	External-tool criterion: a user-supplied command computes the criterion value over the VALUE line protocol
criterion_multinom.hpp	Defines interface for feature selection criterion implementations based on multinomial model
criterion_multinom_bhattacharyya.hpp	Implements Bhattacharyya distance based on multinomial model to serve as feature selection criterion
criterion_negative.hpp	Criterion_Negative returns the negative of another Criterion's result
criterion_normal.hpp	Defines interface for feature selection criterion implementations based on normal model
criterion_normal_bhattacharyya.hpp	Implements Bhattacharyya distance based on normal (gaussian) model to serve as feature selection criterion
criterion_normal_divergence.hpp	Implements Divergence (distance) based on normal (gaussian) model to serve as feature selection criterion
criterion_normal_gmahalanobis.hpp	Implements Generalized Mahalanobis distance based on normal (gaussian) model to serve as feature selection criterion
criterion_subsetsize.hpp	Criterion_Subset_Size returns the (positive) subset size - rates larger subsets higher; wrap in Criterion_Negative to favor smaller subsets
criterion_sumofweights.hpp	Criterion_Sum_Of_Weights returns sum of pre-specified feature weights for features in the evaluated subset
criterion_wrapper.hpp	Criterion_Wrapper adapts Classifier objects to serve as feature selection criterion
criterion_wrapper_bias_estimate.hpp	Wrapper Bias Evaluator - returns classifier bias estimate (difference between accuracy on training and validation data parts)
dense_train_tools.hpp	Shared dense-training helpers: standardized-design build and Cholesky factor/solve
distance.hpp	Defines interface for distance evaluators (to be used primarily in Classifier_kNN)
distance_euclid.hpp	Implements Euclidean (L_2) distance
distance_L1.hpp	Implements L_1 distance
distance_Lp.hpp	Implements L_{p} distance, p equals numerator/denominator (template parameters)
extern_bridge.hpp	Persistent-child-process line-protocol plumbing shared by the external-tool bridges (Classifier_Extern, Criterion_Extern_Value)
indexed_matrix.hpp	Matrix representation and operations, allows operation in selected subspace
indexed_vector.hpp	Vector representation and operations, allows operation in selected subspace
model.hpp	Defines interface for data model implementations
model_multinom.hpp	Implements multinomial model
model_normal.hpp	Implements normal (gaussian) model
regressor.hpp	Defines interface for regressor implementations
regressor_ridge.hpp	Defines Regressor_Ridge — closed-form L2-regularized linear regression for wrapper-based feature selection
regressor_weighted_neighbours.hpp	Experimental regressor, produces weighted average of the predicted variable over neighbours, weighted by distance
wrapper_metrics.hpp	Wrapper criterion metrics beyond accuracy (balanced accuracy, F1, AUC, class weights, regression metrics)
_src_dataio
data_accessor.hpp	Defines interface for data access implementations
data_accessor_splitting.hpp	Defines support for data splitting
data_accessor_splitting_mem.hpp	Implements data access to data cached entirely in memory, concrete file type support is delegated to derived classes
data_accessor_splitting_memARFF.hpp	Implements data access to data cached entirely in memory, read once from ARFF files
data_accessor_splitting_memCSV.hpp	Implements data access to data cached entirely in memory, read once from CSV files
data_accessor_splitting_memTRN.hpp	Implements data access to data cached entirely in memory, read once from TRN files
data_accessor_universe.hpp	Data-free accessor over an abstract feature universe (--features N; pure set-function optimization)
data_file_ARFF.hpp	Enables data accessor objects to read ARFF (Weka) type of data files
data_file_CSV.hpp	Enables data accessor objects to read CSV (delimited text, e.g. Excel export) data files
data_file_TRN.hpp	Enables data accessor objects to read TRN type of data files
data_intervaller.hpp	Container to hold a list of Data_Interval; implements nested interval reduction
data_scaler.hpp	Defines interface for data scaling implementations (for use in data accessors)
data_scaler_to01.hpp	Implements data normalization (of all feature values) to interval [0,1]
data_scaler_void.hpp	Void data scaler, to bypass data normalization
data_scaler_white.hpp	Implements data whitening (as result, each feature mean value is 0 with normalized stddev)
data_splitter.hpp	Defines interface for data splitting implementations (for use in data accessors)
data_splitter_5050.hpp	Implements train/test data splitting: first 50% of original data for training, second 50% for testing
data_splitter_cv.hpp	Implements train/test data splitting: by means of k-fold cross-validation
data_splitter_holdout.hpp	Implements train/test data splitting: use initial x% of original data for training, the rest for testing
data_splitter_leave1out.hpp	Implements train/test data splitting: by means of leave-one-out
data_splitter_randfix.hpp	Implements train/test data splitting: equal to Data_Splitter_RandomRandom except that the same test data is returned in each loop
data_splitter_randrand.hpp	Implements train/test data splitting: use randomly chosen x% of data samples for training and another y% of data for testing, without overlaps
data_splitter_resub.hpp	Implements train/test data splitting: use all data both for training and then once more for testing
_src_global
clonable.hpp	Defines interface for classes that may need to get cloned
error.hpp	FST4 specific error passing class
fst_rand.hpp	FST4-owned pseudo-random number generator (UCRT-compatible LCG), replacing platform-specific std::rand()/std::srand()
global.cpp	Empty compatibility stub — the FST4 library is header-only
global.hpp	Global definitions
subset.hpp	Stores the info on currently selected features, enables generating permutations, etc
thread_pool.hpp	Implements thread scheduler that assigns jobs up to maximum number of threads allowed
_src_search
branch_and_bound_predictor.hpp	Interface for Implementations of Prediction Mechanism in Fast Branch & Bound and Branch & Bound with Partial Prediction
branch_and_bound_predictor_averaging.hpp	Averaging Prediction Mechanism for use in Fast Branch & Bound and Branch & Bound with Partial Prediction
candidate_evaluator.hpp	Abstract candidate-evaluation strategy: the seam through which one search algorithm body evaluates candidates either sequentially in-line or concurrently in a thread pool
candidate_evaluator_sequential.hpp	Implements sequential in-line evaluation of a criterion on a set of subset candidates
candidate_evaluator_threaded.hpp	Implements concurrent evaluation of a criterion on a set of subset candidates
result_tracker.hpp	Defines interface for classes that enable collecting multiple results
result_tracker_dupless.hpp	Enables collecting multiple results
result_tracker_feature_stats.hpp	Computes feature occurrence statistics in a series of evaluated subsets
result_tracker_regularizer.hpp	Enables eventual selection of a different subset
result_tracker_stabileval.hpp	Enables collecting multiple results, then evaluating various stability measures
result_weighter_sampled.hpp	Online per-feature statistics for budgeted (sampled) sequential steps — discounted batch-standardized DAF contrast
search.hpp	Defines interface for search method implementations
search_branch_and_bound.hpp	Implements Branch and Bound template method as basis for more advanced B&B implementations
search_branch_and_bound_basic.hpp	Implements Branch and Bound Basic method, i.e., with randomized node ordering
search_branch_and_bound_fast.hpp	Implements Fast Branch and Bound, i.e., B&B with full utilization of prediction mechanism
search_branch_and_bound_improved.hpp	Implements Improved Branch and Bound (IBB) method, i.e., with fully computed node ordering
search_branch_and_bound_improved_threaded.hpp	DEPRECATED compatibility shim: threaded Improved Branch and Bound is now the merged Search_Branch_And_Bound_Improved with an injected Candidate_Evaluator_Threaded
search_branch_and_bound_partial_prediction.hpp	Implements Branch and Bound with Partial Prediction (BBPP) method, i.e., with predicted node ordering
search_exhaustive.hpp	Implements exhaustive (optimal) search over all feature subsets
search_exhaustive_threaded.hpp	DEPRECATED compatibility shim: threaded exhaustive search is now the merged Search_Exhaustive with an injected Candidate_Evaluator_Threaded
search_monte_carlo.hpp	Repeatedly samples random subsets to eventually yield the one with highest criterion value
search_monte_carlo_threaded.hpp	DEPRECATED compatibility shim: threaded Monte Carlo search is now the merged Search_Monte_Carlo with an injected Candidate_Evaluator_Threaded
search_ranking.hpp	Defines interface for individual feature ranking method implementations
search_ranking_bif.hpp	Implements Best Individual Features, i.e., individual feature ranking
search_ranking_bif_threaded.hpp	DEPRECATED compatibility shim: threaded individual feature ranking is now the merged Search_Ranking_BIF with an injected Candidate_Evaluator_Threaded
search_seq.hpp	Defines interface for sequential search method implementations
search_seq_dos.hpp	Implements Dynamic_Oscillating_Search
search_seq_os.hpp	Implements Oscillating_Search
search_seq_sffs.hpp	Implements Sequential_Forward_Floating_Search and Sequential_Backward_Floating_Search
search_seq_sfrs.hpp	Implements Sequential_Forward_Retreating_Search and Sequential_Backward_Retreating_Search
search_seq_sfs.hpp	Implements Sequential_Forward_Selection and Sequential_Backward_Selection
seq_step.hpp	Defines interface for implementations of one selection step in sequential search type of methods
seq_step_ensemble.hpp	Implements voting ensemble selection step in sequential search type of methods
seq_step_hybrid.hpp	Implements hybrid selection step in sequential search type of methods
seq_step_sampled.hpp	Implements the budgeted (sampled) sequential selection step — per-step candidate evaluations capped, candidates proposed by online per-feature statistics
seq_step_straight.hpp	Implements standard sequential selection step in sequential search type of methods
seq_step_straight_threaded.hpp	DEPRECATED compatibility shim: threaded sequential selection step is now the merged Sequential_Step_Straight with an injected Candidate_Evaluator_Threaded
stopwatch.hpp	Simple tool to measure run time
demos
demo10.cpp	Example 10: Basic Filter-based feature selection
demo52t.cpp	Example 52t: (Threaded SFRS) Result regularization using secondary criterion
demo55.cpp	Example 55: Evaluating Similarity of Two Feature Selection Processes
fst4.cpp	Exposes FST4 functionality as a command-line tool: data file, splitter, scaler, criterion, search algorithm and their settings are selected via command-line arguments