KERNEL (Languedoc-Roussillon Regional grant “chercheur d’avenir)

Massive datas arise from most of the scientific disciplines, including biology, physics, environmental and health science… To handle the huge amount of (hidden) information represented by these datas require a systematic and rigorous approach ranging from databases and data mining to algorithms. A key to improve the algorithmic efficiency is to simplify, to filter, to reduce the given data sets. Underlying most of the heuristic methods or even the exact algorithms, data-preprocessing strategies have been developed since decades. However, only recently an attempt to formalize the concept of preprocessing algorithm has been proposed through the kernelization theory. To date, kernelization algorithms have been mostly designed to tackle NP-hard graph algorithmic problems. The objective of the KERNEL project is to expand to domain of application of the kernelization theory to other fields at the interface with computer science.

Dates: 2012-2015

Project leader: C. Paul (LIRMM, CNRS, Université Montpellier)