Hitachi Vantara Pentaho Community Wiki
Child pages
  • PrincipalComponents (filter)
Skip to end of metadata
Go to start of metadata

Package

weka.filters.unsupervised.attribute

Synopsis

Performs a principal components analysis and transformation of the data.
Dimensionality reduction is accomplished by choosing enough eigenvectors to account for some percentage of the variance in the original data – default 0.95 (95%).
Based on code of the attribute selection scheme 'PrincipalComponents' by Mark Hall and Gabi Schmidberger.

Options

The table below describes the options available for PrincipalComponents.

Option

Description

maximumAttributeNames

The maximum number of attributes to include in transformed attribute names.

maximumAttributes

The maximum number of PC attributes to retain.

normalize

Normalize input data.

varianceCovered

Retain enough PC attributes to account for this proportion of variance.

Capabilities

The table below describes the capabilites of PrincipalComponents.

Capability

Supported

Class

Nominal class, Numeric class, Missing class values, Binary class, Date class, No class

Attributes

Numeric attributes, Nominal attributes, Unary attributes, Binary attributes, Empty nominal attributes, Date attributes, Missing values

Min # of instances

0

  • No labels