bimodality_coefficient

Should the finite sample size correction be applied to the
skewness and kurtosis measures?  Defaults to TRUE.

finite


This function calculates the Bimodality Coefficient of a data vector
with the option for a finite sample (bias) correction. This bias
correction is important to correct for the (well-documented) 
finite sample bias. 
The bimodality coefficient has a range of zero to one (that is: [0,1])
where a value greater than "5/9" suggests bimodality. The maximum value
of one ("1") can only be reached when the distribution is composed of
two point masses.


nonparametric

Measure

Bimodality

Modality

Designed with a dual purpose of
accurately estimating the mode (or modes) as well as characterizing
the modality of data. The specific application area includes complex
or mixture distributions particularly in a big data environment.
The heterogeneous nature of (big) data may require deep introspective
statistical and machine learning techniques, but these statistical tools
often fail when applied without first understanding the data. In small
datasets, this often isn't a big issue, but when dealing with large scale
data analysis or big data thoroughly inspecting each dimension
typically yields an O(n^n-1) problem. As such, dealing with big data
require an alternative toolkit. This package not only identifies the
mode or modes for various data types, it also provides a programmatic
way of understanding the modality (i.e. unimodal, bimodal, etc.) of
a dataset (whether it's big data or not). See
<http://www.sdeevi.com/modes_package> for examples and discussion.

bimodality_coefficient: Bimodality Coefficient

Description

Usage

Arguments

References

Examples