Whitepaper: Technique for robust data-driven clustering
How does data-driven targeting work?
This paper describes an algorithm designed to reliably cluster real data of high dimension into self-similar groups. The paper was written by the authors of the algorithm and covers at a high level the processes and reasoning that gave rise to the approach taken. The intention is to describe some of the common problems encountered in this style of analysis and the characteristics of the solutions invented or adopted to cope with them. The algorithm is embodied in a software program written in Java. Specifics of the implementation of the algorithm are not covered in this paper.
The authors have developed the algorithm and associated software for exclusive use by Clusters, the segmentation specialists.
What does the algorithm do?
In the current implementation of the algorithm there are four steps:
Want to read the full whitepaper?
Get the full whitepaper by filling in the form below.