Skip to main content

Relevance Vector Machine

Synopsis

This operator is an implementation of Relevance Vector Machine (RVM) which is a probabilistic method both for classification and regression.

Description

The Relevance Vector Machine operator is a probabilistic method both for classification and regression. The implementation of the relevance vector machine is based on the original algorithm described by 'Tipping/2001'. The fast version of the marginal likelihood maximization ('Tipping/Faul/2003') is also available if thervm typeparameter is set to 'Constructive-Regression-RVM'.

A Relevance Vector Machine (RVM) is a machine learning technique that uses Bayesian inference to obtain parsimonious solutions for regression and classification. The RVM has an identical functional form to the support vector machine, but provides probabilistic classification. It is actually equivalent to a Gaussian process model with a certain covariance function. Compared to that of support vector machines (SVM), the Bayesian formulation of the RVM avoids the set of free parameters of the SVM (that usually require cross-validation-based post-optimizations). However RVMs use an expectation maximization (EM)-like learning method and are therefore at risk of local minima. This is unlike the standard sequential minimal optimization(SMO)-based algorithms employed by SVMs, which are guaranteed to find a global optimum.

Input

training set

This input port expects an ExampleSet. This operator cannot handle nominal attributes; it can be applied on data sets with numeric attributes. Thus often you may have to use the Nominal to Numerical operator before the application of this operator.

Output

model

The RVM is applied and the resultant model is delivered from this output port. This model can now be applied on unseen data sets.

example set

The ExampleSet that was given as input is passed without changing to the output through this port. This is usually used to reuse the same ExampleSet in further operators or to view the ExampleSet in the Results Workspace.

Parameters

Rvm type

This parameter specifies the type of RVM Regression. The following options are available: Regression-RVM, Classification-RVM and Constructive-Regression-RVM.

Kernel type

The type of the kernel function is selected through this parameter. Following kernel types are supported:rbf, cauchy, laplace, poly, sigmoid, Epanechnikov, gaussian combination, multiquadric

Kernel lengthscale

This parameter specifies the lengthscale to be used in all kernels.

Kernel degree

这是内核parameter degree. This is only available when thekernel typeparameter is set topolynomialorepachnenikov.

Kernel bias

This parameter specifies the bias to be used in the poly kernel.

Kernel sigma1

这是内核parameter sigma1. This is only available when thekernel typeparameter is set toepachnenikov,gaussian combinationormultiquadric.

Kernel sigma2

这是内核parameter sigma2. This is only available when thekernel typeparameter is set togaussian combination.

Kernel sigma3

这是内核parameter sigma3. This is only available when thekernel typeparameter is set togaussian combination.

Kernel shift

这是内核parameter shift. This is only available when thekernel typeparameter is set tomultiquadric.

Kernel a

这是内核parameter a. This is only available when thekernel typeparameter is set tosigmoid

Kernel b

这是内核parameter b. This is only available when thekernel typeparameter is set tosigmoid

Max iteration

This parameter specifies the maximum number of iterations to be used.

Min delta log alpha

The iteration is aborted if the largest log alpha change is smaller thanmin delta log alpha.

Alpha max

The basis function is pruned if its alpha is larger than thealpha max.

Use local random seed

This parameter indicates if alocal random seedshould be used for randomization. Using the same value oflocal random seedwill produce the same randomization.

Local random seed

This parameter specifies thelocal random seed. This parameter is only available if theuse local random seedparameter is set to true.