demo22.cpp File Reference

Example 22: Randomized feature selection with Oscillating Search. More...

#include <boost/smart_ptr.hpp>
#include <exception>
#include <iostream>
#include <cstdlib>
#include <string>
#include <vector>
#include "error.hpp"
#include "global.hpp"
#include "subset.hpp"
#include "data_intervaller.hpp"
#include "data_splitter.hpp"
#include "data_splitter_randrand.hpp"
#include "data_scaler.hpp"
#include "data_scaler_void.hpp"
#include "data_accessor_splitting_memTRN.hpp"
#include "data_accessor_splitting_memARFF.hpp"
#include "criterion_normal_bhattacharyya.hpp"
#include "criterion_wrapper.hpp"
#include "distance_euclid.hpp"
#include "classifier_knn.hpp"
#include "seq_step_straight.hpp"
#include "search_seq_os.hpp"

Include dependency graph for demo22.cpp:

Functions
int	main ()

---

Detailed Description

Example 22: Randomized feature selection with Oscillating Search.

---

Function Documentation

int main

(

)

Example 22: Randomized feature selection with Oscillating Search.

Selects features using randomized OS algorithm and normal Bhattacharyya distance as feature selection criterion. Target subset size must be set by user because a) Bhattacharyya is monotonous (global optimization would yield full set of features) and b) OS requires initial subset of target cardinality. OS is called here repeatedly from various random initial points as long as there comes no improvement during consecutive 5 runs. This mechanism considerably improves chances to avoid local maxima. Bhattacharyya is evaluated on randomly chosen 75% of data samples. The resulting feature subset is eventually validated by means of 3NN classification accuracy estimation on on the remaining 25% of data samples.

References FST::Search_OS< RETURNTYPE, DIMTYPE, SUBSET, CRITERION, EVALUATOR >::search(), and FST::Search_OS< RETURNTYPE, DIMTYPE, SUBSET, CRITERION, EVALUATOR >::set_delta().