Affordable Supercomputing for Data Mining Applications
András A. Benczúr · Procedia Computer Science · 2011
Business intelligence, e-science and Web mining are rapidly growing sources of extreme large scale problems. We demonstrate that emerging means of supercomputing-for-the-masses, distributed or many-core architectures may provide feasible and affordable solution to many applications. We give a brief overview in four areas: • Web data processing to federate Future Internet Research (FIRE) facilities in a recently started project; • Image retrieval over tasks of the PASCAL and PROMISE networks of excellence; • Efficiency of distributed data warehouses for log processing and entity resolution; • Network analysis, visualization and navigation, in the VAST Challenges and European security projects. We pass beyond existing technologies in implementing new frameworks, combining distributed and multicore technologies, and seeking new, breakthrough applications in an interdisciplinary research with application partners.