pairwise

Documentation

Overview

Package pairwise implements utilities to evaluate pairwise distances or inner product (via kernel).

Index

• type Chebyshev
  • func NewChebyshev() *Chebyshev
  • func (c *Chebyshev) Distance(vectorX *mat.Dense, vectorY *mat.Dense) float64
• type Cosine
  • func NewCosine() *Cosine
  • func (c *Cosine) Distance(vectorX *mat.Dense, vectorY *mat.Dense) float64
  • func (c *Cosine) Dot(vectorX *mat.Dense, vectorY *mat.Dense) float64
• type Cranberra
  • func NewCranberra() *Cranberra
  • func (c *Cranberra) Distance(vectorX *mat.Dense, vectorY *mat.Dense) float64
• type Euclidean
  • func NewEuclidean() *Euclidean
  • func (e *Euclidean) Distance(vectorX *mat.Dense, vectorY *mat.Dense) float64
  • func (e *Euclidean) InnerProduct(vectorX *mat.Dense, vectorY *mat.Dense) float64
• type Manhattan
  • func NewManhattan() *Manhattan
  • func (m *Manhattan) Distance(vectorX *mat.Dense, vectorY *mat.Dense) float64
• type PairwiseDistanceFunc
• type PolyKernel
  • func NewPolyKernel(degree int) *PolyKernel
  • func (p *PolyKernel) Distance(vectorX *mat.Dense, vectorY *mat.Dense) float64
  • func (p *PolyKernel) InnerProduct(vectorX *mat.Dense, vectorY *mat.Dense) float64
• type RBFKernel
  • func NewRBFKernel(gamma float64) *RBFKernel
  • func (r *RBFKernel) InnerProduct(vectorX *mat.Dense, vectorY *mat.Dense) float64

Types

type Chebyshev

type Chebyshev struct{}

func NewChebyshev

func NewChebyshev() *Chebyshev

func (*Chebyshev) Distance

func (c *Chebyshev) Distance(vectorX *mat.Dense, vectorY *mat.Dense) float64

type Cosine

type Cosine struct{}

func NewCosine

func NewCosine() *Cosine

func (*Cosine) Distance

func (c *Cosine) Distance(vectorX *mat.Dense, vectorY *mat.Dense) float64

Distance computes Cosine distance. It will return distance which represented as 1-cos() (ranged from 0 to 2).

func (*Cosine) Dot

func (c *Cosine) Dot(vectorX *mat.Dense, vectorY *mat.Dense) float64

Dot computes dot value of vectorX and vectorY.

type Cranberra

type Cranberra struct{}

func NewCranberra

func NewCranberra() *Cranberra

func (*Cranberra) Distance

func (c *Cranberra) Distance(vectorX *mat.Dense, vectorY *mat.Dense) float64

type Euclidean

type Euclidean struct{}

func NewEuclidean

func NewEuclidean() *Euclidean

func (*Euclidean) Distance

func (e *Euclidean) Distance(vectorX *mat.Dense, vectorY *mat.Dense) float64

Distance computes Euclidean distance (also known as L2 distance).

func (*Euclidean) InnerProduct

func (e *Euclidean) InnerProduct(vectorX *mat.Dense, vectorY *mat.Dense) float64

InnerProduct computes a Eucledian inner product.

type Manhattan

type Manhattan struct{}

func NewManhattan

func NewManhattan() *Manhattan

func (*Manhattan) Distance

func (m *Manhattan) Distance(vectorX *mat.Dense, vectorY *mat.Dense) float64

Distance computes the Manhattan distance, also known as L1 distance. == the sum of the absolute values of elements.

type PairwiseDistanceFunc

type PairwiseDistanceFunc interface {
	Distance(vectorX *mat.Dense, vectorY *mat.Dense) float64
}

type PolyKernel

type PolyKernel struct {
	// contains filtered or unexported fields
}

func NewPolyKernel

func NewPolyKernel(degree int) *PolyKernel

NewPolyKernel returns a d-degree polynomial kernel

func (*PolyKernel) Distance

func (p *PolyKernel) Distance(vectorX *mat.Dense, vectorY *mat.Dense) float64

Distance computes distance under the polynomial kernel (maybe not needed?)

func (*PolyKernel) InnerProduct

func (p *PolyKernel) InnerProduct(vectorX *mat.Dense, vectorY *mat.Dense) float64

InnerProduct computes the inner product through a kernel trick K(x, y) = (x^T y + 1)^d

type RBFKernel

type RBFKernel struct {
	// contains filtered or unexported fields
}

func NewRBFKernel

func NewRBFKernel(gamma float64) *RBFKernel

NewRBFKernel returns a representation of a Radial Basis Function Kernel

func (*RBFKernel) InnerProduct

func (r *RBFKernel) InnerProduct(vectorX *mat.Dense, vectorY *mat.Dense) float64

InnerProduct computes the inner product through a kernel trick K(x, y) = exp(-gamma * ||x - y||^2)