mltk::classifier::KNNClassifier< T, Callable > Class Template Reference

Wrapper for the implementation of the K-Nearest Neighbors classifier algorithm. More...

#include <KNNClassifier.hpp>

Inheritance diagram for mltk::classifier::KNNClassifier< T, Callable >:

Collaboration diagram for mltk::classifier::KNNClassifier< T, Callable >:

Public Member Functions
	KNNClassifier (size_t _k, std::string _algorithm="brute")
	KNNClassifier (Data< T > &_samples, size_t _k, std::string _algorithm="brute")
bool	train () override
	Function that execute the training phase of a Learner. More...
double	evaluate (const Point< T > &p, bool raw_value=false) override
	Returns the class of a feature point based on the trained Learner. More...
Callable &	metric ()
void	setPrecomputedDistances (metrics::dist::BaseMatrix _distances)
metrics::dist::DistanceMatrix< Callable >	precomputeDistances (mltk::Data< T > &data, bool diagonal=false, const size_t threads=std::thread::hardware_concurrency())
Public Member Functions inherited from mltk::classifier::PrimalClassifier< double >
	PrimalClassifier (DataPointer< double > samples)
	PrimalClassifier (mltk::Data< double > &samples)
	PrimalClassifier (const PrimalClassifier< double > &primal_learner)
double	evaluate (const Point< double > &p, bool raw_value=false) override
	Returns the class of a feature point based on the trained Learner. More...
std::string	getFormulationString () override
	getFormulationString Returns a string that represents the formulation of the learner (Primal or Dual). More...
double	getP () const
	GetP Return the value of the p norm. More...
double	getQ () const
	GetQ Return the value of the q norm. More...
void	setqNorm (double q)
	setqNorm Set the q norm used by the classifier. (Euclidean norm is the default) More...
void	setpNorm (double p)
	setpNorm Set the p norm used by the classifier. (Euclidean norm is the default) More...
void	setFlexible (double flexible)
	Set flexibity of the classifier. More...
void	setAlphaAprox (double alpha_aprox)
	Set the percentage of the aproximation. More...
Public Member Functions inherited from mltk::classifier::Classifier< T >
	Classifier (DataPointer< T > samples)
	Classifier (const Classifier< T > &classifier)
virtual double	evaluateProbability (const mltk::Point< double > &p)
mltk::Point< int >	batchEvaluateProbability (const mltk::Data< T > &data)
Solution	getSolution () const
	getSolution Returns the solution of the classifier. More...
Solution *	getSolutionRef ()
	getSolution Returns a reference to the solution of the classifier. More...
void	setGamma (double gamma)
	Set the gamma (margin) of the classifier. More...
void	setW (std::vector< double > w)
	setW Set the weights vector of the classifier. More...
void	setSolution (Solution solution)
	setSolution Set a solution for the classifier. More...
Public Member Functions inherited from mltk::Learner< T >
	Learner (DataPointer< T > _samples)
	Learner (const Learner< T > &learner)
virtual mltk::Point< double >	batchEvaluate (const Data< T > &data)
	evaluate a batch of points. More...
auto	getSamples ()
	Get the Data used by the learner. More...
double	getElapsedTime () const
	Get the elapsed time in the training phase of the Learner. More...
int	getCtot () const
	Get the total number of updates of the Learner. More...
int	getSteps () const
	getSteps Returns the number of steps through the data by the Learner. More...
int	getUpdates () const
	getUpdates Returns the number of updates needed to get to the the solution. More...
double	getMaxTime () const
	getMaxTime Returns the maximum running time in the training phase of the Learner. More...
double	getPredictionProbability () const
	Get the probability of the last prediction. More...
void	setSeed (const size_t _seed)
	Set the seed to be used by the learner. More...
virtual void	setSamples (const Data< T > &data)
	setSamples Set the samples used by the Learner. More...
virtual void	setSamples (DataPointer< T > data)
	setSamples Set the samples used by the Learner. More...
void	setTimer (Timer _timer)
	setTimer Set the timer used by the Learner. More...
void	setSteps (int _steps)
	Set the partial number of steps used in the training phase of the Learner. More...
void	setCtot (int _ctot)
	Set the partial number of updates of the Learner. More...
void	setVerbose (int _verbose)
	Set the level of verbose. More...
void	setStartTime (double stime)
	setStartTime Set the initial time of the Learner. More...
void	setMaxTime (double maxtime)
	Set the max time of execution. More...
void	setEPS (double eps)
	setEPS Set the precision of the Learner. More...
void	setMaxIterations (int max_it)
	setMaxIterations Set the max number of iterations of the Learner. More...
void	setMaxEpochs (int MAX_EPOCHS)
	Set the max number of epochs for the learner training. More...
void	setMaxUpdates (int max_up)
	setMaxIterations Set the max number of updates of the Learner. More...
void	setLearningRate (double learning_rate)
	Set the learning rate of the Learner. More...

Additional Inherited Members
Protected Attributes inherited from mltk::classifier::PrimalClassifier< double >
std::vector< double >	w
	Weights vector. More...
double	q
	Norm used in the classification. (Euclidean Norm is the default) More...
double	p
double	flexible
	Flexibility. More...
double	alpha_aprox
	Percentage of aproximation of the result. More...
Protected Attributes inherited from mltk::classifier::Classifier< T >
bool	hasInitialSolution = false
	Inform if there's an initial solution. More...
std::vector< Point< T > >	svs
	Support vectors points. More...
Solution	solution
	Classifier solution. More...
double	gamma = 0
	Classifier margin. More...
Protected Attributes inherited from mltk::Learner< T >
std::shared_ptr< Data< T > >	samples
	Samples used in the model training. More...
double	rate
	Learning rate. More...
double	start_time
	Initial time. More...
double	max_time
	Maximum time of training. More...
int	steps
	Number of steps in the data. More...
int	ctot
	Number of updates of the weights. More...
double	EPS
	Max precision. More...
double	MIN_INC
	Minimun Increment. More...
int	MAX_IT
	Max number of iterations. More...
int	MAX_UP
	Max number of updates. More...
int	MAX_EPOCH
int	verbose
	Verbose level of the output. More...
Timer	timer
	Timer used to measure the time elapsed in the execution of a Learner. More...
size_t	seed
	seed for random operations. More...
double	pred_prob

Detailed Description

template<typename T = double, typename Callable = metrics::dist::Euclidean<T>>
class mltk::classifier::KNNClassifier< T, Callable >

Wrapper for the implementation of the K-Nearest Neighbors classifier algorithm.

Member Function Documentation

evaluate()

template<typename T , typename Callable >

double mltk::classifier::KNNClassifier< T, Callable >::evaluate ( const Point< T > &  p, bool  raw_value = false  )

overridevirtual

Returns the class of a feature point based on the trained Learner.

Parameters

p	Point to be evaluated.

Returns

Prediction made by the learner.

Implements mltk::Learner< T >.

train()

template<typename T , typename Callable >

bool mltk::classifier::KNNClassifier< T, Callable >::train

overridevirtual

Function that execute the training phase of a Learner.

Returns

void

Implements mltk::Learner< T >.

---

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

• KNNClassifier.hpp