RapidMiner (formerly YALE) supports the design and evaluation of data mining processes. That means that you can interactively design such processes and inspect them until the desired results are achieved.

All data mining processes are designed as operator trees. Unlike most other Data Mining suits, the operators in RapidMiner are not defined in a graph layout where components are positioned and connected by the user. The data flow is always the same and follows the well known concept of a depth first search. This drastically eases the design of data mining processes. Child operators are invoked by their parents and work on the data their parents provide. This also allows for loops which are controlled by the parent operators. The trees are defined in XML. This turns RapidMiner into a powerful scripting language engine for data mining experiments and together with the graphical user interface into a first and complete IDE for Knowledge Discovery.

After the design phase, the process is executed and evaluated. The modular design of RapidMiner simply allows process changes before re-evaluation. The open-source data mining software RapidMiner has demonstrated this strength in many real-world applications ranging from text mining, data stream mining and stock prediction to multimedia mining.