Elements of an agile discovery environment

Abstract

Machine learning methods and data mining techniques have proved to be quite helpful in a number of discovery tasks. However, the most popular modern tools in this area do not tend to back the discovery process properly. In this paper we investigate the reasons that prevent modern data mining tools from becoming convenient and productive discovery environments. We come up with principles of an agile discovery environment, i.e. a data mining-driven software designed to support the process of discovery. © Springer-Verlag Berlin Heidelberg 2003.