Icon

R Graphics - use R package modes to detect bimodality, calculate coefficients and create a chart

use R package 'modes' to detect bimodality, calculate coefficients and create a chart

You could use that in KNIME and see if you can use the statistics that are provided by this package to make a decision. It also creates a plot where you can visually inspect if there is a Bi-Modal distribution. You might adapt that to test for further distributions.

Please note I am not an expert in these statistics, just built them into a workflow :slight_smile:. The description 9 for the package for example states that for the:

bimodality_coefficient - "The bimodality coefficient has a range of zero to one (that is: [0,1]) where a value greater than “5/9” suggests bimodality. "

So with 0.774 being larger than 0.556 the statistic here would indicate that the distribution is bimodal. And the visual inspection seems to support that.

I was toying around with creating a variable to bring the description into the graphic but gave up for now. You might include that ins some future graphic you might create.

URL: forum entry https://forum.knime.com/t/identify-bi-modal-distribution/13269/3?u=mlauber71
URL: Find the Modes and Assess the Modality of Complex and Mixture Distributions, Especially with Big Datasets https://github.com/sathish-deevi/modes-Package/tree/master
URL: modes-Package, An R package that calculates various mode and modal measures for complex distributions and big data https://github.com/sathish-deevi/modes-Package/tree/master
URL: Hub: more R Graphics with KNIME https://hub.knime.com/search?type=Workflow&tag=r,graphic&sort=best
URL: Medium Blog: KNIME and R — installation across operating systems — some remarks https://medium.com/p/6494a2a498cc
URL: Medium Blog: Exploring the Power of R Graphics with KNIME: A Collection of Examples https://medium.com/p/9241e033e4ac

Nodes

Extensions

Links