Application of Data Mining Technologies to Tokamak Plasma

70276 This project will develop software tools that can be used by experimental fusion scientists to develop empirical physics models from large quantities of measured data. In particular, powerful data mining algorithms will be developed to extract, directly from data, several kinds of information that support the manner in which fusion scientists represent knowledge. Ultimately, the algorithms will be wrapped into a data mining toolset or environment with general applicability. Phase I will implement an algorithm that derives continuous models from large quantities of data. The algorithm will be applied to a specific plasma physics problem to demonstrate its utility. Commercial Applications and Other Benefits as described by the awardee: In addition to fusion science, the software tools should be applicable to many other scientific problems. In some industries, the recognized need for data mining tools has already outstripped current technology. In addition, many industries that have natural applications for these tools are not yet aware that this technology exists.