Design of efficient Input/output intensive data mining applications

The goal of Data Mining is to discover knowledge hidden in data repositories. This activity has recently attracted a lot of attention. Because of the huge datasets to be accessed, most data mining algorithms must consider the I/O actions carefully, hiding or minimizing their effects. At CNUCE-CNR we are studying how the architectural features of modern high performance computers and operating systems can be exploited wwhen designing efficient data mining algorithms that can cope with huge datasets