Jump to content

2 Screenshots

Summary

This TERR data function and Spotfire dxp file analyzes your data and creates a self-organizing map.

Overview

A self-organizing map is a form of unsupervised learning and can provide a high level view of patterns that exist in a set of data.

The result is shown as a two-dimensional grid that can be regarded as a series of cells, or buckets - every data point will fall into one of these buckets.  The mapping is continuous so the buckets will vary smoothly across the surface.   The surface can be sampled finely or coarsely; and can be sampled in a rectangular or hexagonal arrangement.   The user can control both of these parameters.

The dxp displays the map as a series of "star" patterns, with each arm of the star corresponding to each of the variables chosen.  The arms are scaled to geometrically fit on the pattern, with short and long arms corresponding to small and large values of the original variable.

Functionally, a self-organizing map is similar to methods like kmeans clustering, except that kmeans clustering gives the user the clusters, whereas here the user can explore the groupings at will.

Using this Spotfire dxp file you can analyze your own data and discover new patterns and insights using the self organizing map.

 

User's guide

Much more info and details about this functionality can be found on this community article.

 

Data function library

There exists a large number of data functions covering various features. Feel free to review what is available on the Data Function Library.

 

Release 1.0.0

Published: August 2020

Initial release


×
×
  • Create New...