Using the well-known minimal spanning tree algorithm by Kruskal, we have developed a clustering tool that organises data based on the similarities of individual elements and groups of data items.

This technology has been used in a project where equity shares traded on the JSE were clustered in groups according to their relative price and volume movements. This project was aimed at analysing groups of similarly performing equities based on their daily, monthly and quarterly volatilities.

The question that this technology answers is: "Are there any significant and recurring pattern groupings that exist in my dataset, and if so, which of these patterns could be grouped together?"

Similarity clustering can be further applied in marketing surveys, where grouping participants according to demographics, income and other factors, serves to make the target market more easily segmented and analysable.