Data Mining and Mathematical Programming

October 10-13, 2006
Centre de recherches mathématiques

Organizers: Pierre Hansen (HÉC Montréal) and
Panos Pardalos (Florida)

Data mining is a fast-growing discipline that uses techniques from several subfields of applied mathematics, including operations research and statistics. Most data mining techniques fall into one of the following categories : predictive modelling, clustering, dependency modelling, data summarization, and change and deviation detection (see Mathematical Programming for Data Mining : Formulations and Challenges, by Bradley, Fayyad and Mangasarian). Many of the problems that arise in these various techniques can be formulated as mathematical programming problems. This workshop will feature applications of exact or heuristic algorithms for solving mathematical programs (linear or nonlinear, convex or nonconvex) to the fundamental problems in data mining, in particular clustering, discrimination and search for relations.