`dataFrame.columns.stat — divergence()`

Description

The divergence() method of the stat object computes the divergence between the distributions represented by two selected columns.

Signature

dataFrame.columns(...columnNames).stat.divergence({ method: 'jensenShannon' })

Scope: columns
Family: stat
Returns: number

Arguments

...columnNames ( string[] ): The name of the columns from which to compute divergence.
options (object): Divergence computation options.

Option

method (string)

The divergence measure used for the computation.

jensenShannon (default)
kullbackLeibler

Returns

divergence (number): The computed divergence value.

Notes

The method requires exactly two selected columns.
The columns are interpreted as distributions.
Kullback-Leibler divergence is directional and depends on the order of the selected columns.
Jensen-Shannon divergence is symmetric and does not depend on column order.
A divergence of 0 indicates identical distributions.
Larger values indicate greater differences between distributions.
When using Kullback-Leibler divergence, reversing the column order may produce a different result.

Example

// evaluate the divergence of the values of 2 columns of the dataFrame
var divergence = dataFrame.columns('age', 'income').stat.divergence();

// log the stat
notebook.log(divergence);