realKD.org is dedicated to increasing the real-world impact of academic research on pattern discovery and other data mining techniques. In data mining research groups all around the world, hundreds of brilliant ideas are developed every year—ideas that could help real users to solve real problems with real data. However, only a small part of those ideas end up published in one of the few highly competitive scientific conferences or journals. Moreover, even of those ideas which reach scientific acceptance, only a small fraction is implemented in a robust and reusable form that would allow them to develop a lasting impact in different application scenarios.
This web site seeks to help data mining researchers to increase the impact of their ideas on both levels: to make them accessible for solving real problems as well as to reach their academic acceptance in the first place. To this end, we endorse two free open source Java projects and provide integration and usage support. One of them, realKD, is an architectural framework for the easy implementation of data mining algorithm that is accessible for real users. The other one, Creedo, enables the definition of visual data analysis systems based on realKD’s algorithms that users can access easily through a web browser. Building on this capability, Creedo allows researchers to define and run experimental studies with real users to support the value of their contributions.