cmp::classifier::KNN Class Reference

K-Nearest Neighbors classifier helper class. More...

#include <classifier.h>

Public Member Functions
void	compute (const std::vector< Eigen::VectorXd > &points, size_t k)
	Computes the nearest neighbor index map for the given points.
const std::vector< size_t > &	operator[] (size_t i) const
	Accesses the list of nearest neighbor indices for the i-th point.
size_t	nPoints () const
	Returns the total number of points in the dataset.
size_t	k () const
	Returns the value of k (number of neighbors).

Private Attributes
std::vector< std::vector< size_t > >	neighbours_
	Cache containing nearest neighbor indices for each data point.
size_t	kNearestValue_
	The number of nearest neighbors (k).
size_t	nPoints_
	Total number of data points.

Detailed Description

K-Nearest Neighbors classifier helper class.

Mathematical Formulation For a query point \(\mathbf{x}\), K-Nearest Neighbors identifies the set of \(k\) closest points \(\mathcal{N}_k(\mathbf{x})\) in the reference dataset \(\mathcal{D} = \{\mathbf{x}_i\}_{i=1}^N\) under a chosen distance metric \(d(\mathbf{x}, \mathbf{x}_i)\) (typically Euclidean distance):

\[ d(\mathbf{x}, \mathbf{x}_i) = \|\mathbf{x} - \mathbf{x}_i\|_2 \]

Implementation Algorithm The compute method builds an indexing/neighbour structure. For each point in the dataset, it searches and stores the indices of its \(k\) nearest neighbors.

Member Function Documentation

compute()

void cmp::classifier::KNN::compute	(	const std::vector< Eigen::VectorXd > &	points,
		size_t	k
	)

Computes the nearest neighbor index map for the given points.

Parameters

points	Vector of data points.
k	Number of nearest neighbors to compute.

k()

size_t cmp::classifier::KNN::k

(

)

const

Returns the value of k (number of neighbors).

Returns

Number of neighbors.

nPoints()

size_t cmp::classifier::KNN::nPoints

(

)

const

Returns the total number of points in the dataset.

Returns

Size of the dataset.

operator[]()

const std::vector< size_t > & cmp::classifier::KNN::operator[]

(

size_t

i

)

const

Accesses the list of nearest neighbor indices for the i-th point.

Parameters

i	Index of the query point.

Returns

Reference to a vector of neighbor indices.

Member Data Documentation

kNearestValue_

size_t cmp::classifier::KNN::kNearestValue_

private

The number of nearest neighbors (k).

neighbours_

std::vector<std::vector<size_t> > cmp::classifier::KNN::neighbours_

private

Cache containing nearest neighbor indices for each data point.

nPoints_

size_t cmp::classifier::KNN::nPoints_

private

Total number of data points.

---

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

• include/classifier.h
• src/classifier.cpp