Package

weka.filters.unsupervised.attribute

Synopsis

Converts the propositional instance dataset into multi-instance dataset (with relational attribute). When normalize or standardize a multi-instance dataset, a MIToSingleInstance filter can be applied first to convert the multi-instance dataset into propositional instance dataset. After normalization or standardization, may use this PropositionalToMultiInstance filter to convert the data back to multi-instance format.

Note: the first attribute of the original propositional instance dataset must be a nominal attribute which is expected to be bagId attribute.

Options

The table below describes the options available for PropositionalToMultiInstance.

Option	Description
randomize	Whether the order of the generated data is randomized.
seed	The random seed used by the random number generator

Capabilities

The table below describes the capabilites of PropositionalToMultiInstance.

Capability	Supported
Class	Date class, Numeric class, Unary class, Binary class, String class, Missing class values, Nominal class, Relational class, Empty nominal class, No class
Attributes	Binary attributes, Nominal attributes, Numeric attributes, Unary attributes, String attributes, Empty nominal attributes, Missing values, Date attributes
Min # of instances	0