Hitachi Vantara Pentaho Community Wiki
Child pages
  • StratifiedRemoveFolds
Skip to end of metadata
Go to start of metadata

Package

weka.filters.supervised.instance

Synopsis

This filter takes a dataset and outputs a specified fold for cross validation. If you do not want the folds to be stratified use the unsupervised version.

Options

The table below describes the options available for StratifiedRemoveFolds.

Option

Description

fold

The fold which is selected.

invertSelection

Whether to invert the selection.

numFolds

The number of folds to split the dataset into.

seed

the random number seed for shuffling the dataset. If seed is negative, shuffling will not be performed.

Capabilities

The table below describes the capabilites of StratifiedRemoveFolds.

Capability

Supported

Class

Numeric class, Empty nominal class, Nominal class, Date class, Relational class, Binary class, String class, Unary class, Missing class values

Attributes

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

Min # of instances

0

  • No labels