As a guest user you are not logged in or recognized by your IP address. You have
access to the Front Matter, Abstracts, Author Index, Subject Index and the full
text of Open Access publications.
Today data mining is more and more extensively used by very competitive enterprises. This development, brought by the increasing availability of massive datasets, is only possible if solutions to challenges, both theoretic and operational, are found: identify algorithms which can be used to produce models when datasets have thousands of variables and millions of observations; learn how to run and control the correct execution of hundreds of models; automate the data mining process. We will present these constraints in industrial contexts; we will show how to exploit theoretical results (coming from Vladimir Vapnik's work) to produce robust models; we will give a few examples of real-life applications. We will thus demonstrate that it is indeed possible to industrialize data mining so as to turn it into an easy-to-use component whenever data is available.