Design and implementation of an environment for frequent pattern mining on multi core architectures.

Benjamin Négrevergne
under the direction of Marie-Christine Rousset and Alexandre Termier.
Funding: MENRT (Sep 08 - Sep 11)

Last years the community of data-mining algorithms users has shown more and more interest in the extraction of complex patterns such as sequences rees or graphs, in huge data sets. To output the results in a reasonable amount of time the new algorithms need more omputing power. However it is known by the processor's designers that the processor's power will not increase in terms of frequency any ore, but in number of cores. To exploit this new form of
computing power the new algorithms must be parallel algorithms. With this thesis we intend to provide to the data-mining community the tools hey needs to write efficient and parallel algorithms.