MLPACK  1.0.4
Classes
mlpack::kernel Namespace Reference

Kernel functions. More...

Classes

class  CosineDistance
 The cosine distance (or cosine similarity). More...
class  EpanechnikovKernel
 The Epanechnikov kernel, defined as. More...
class  ExampleKernel
 An example kernel function. More...
class  GaussianKernel
 The standard Gaussian kernel. More...
class  HyperbolicTangentKernel
 Hyperbolic tangent kernel. More...
class  LaplacianKernel
 The standard Laplacian kernel. More...
class  LinearKernel
 The simple linear kernel (dot product). More...
class  PolynomialKernel
 The simple polynomial kernel. More...
class  PSpectrumStringKernel
 The p-spectrum string kernel. More...
class  SphericalKernel
class  TriangularKernel
 The trivially simple triangular kernel, defined by. More...

Detailed Description

Kernel functions.

This namespace contains kernel functions, which evaluate some kernel function $ K(x, y) $ for some arbitrary vectors $ x $ and $ y $ of the same dimension. The single restriction on the function $ K(x, y) $ is that it must satisfy Mercer's condition:

\[ \int \int K(x, y) g(x) g(y) dx dy \ge 0 \]

for all square integrable functions $ g(x) $.

The kernels in this namespace all implement the same methods as the ExampleKernel class. Any additional custom kernels should implement all the methods that class implements; in addition, any method using a kernel should rely on any arbitrary kernel function class having a default constructor and a function

 double Evaluate(arma::vec&, arma::vec&);