Home page

Photos

Aims

Publication in Journals

Program

Proceedings

Important dates

Registration

Keynote Speakers

Invited sessions

Submissions

Instructions to Authors

Call For Papers

IBM Award

Solving Problems

Organisation Committee

Program Committee

Sponsors

Venue

Accommodation

Contact us

 


Keynote Speakers


-  Giovani Barone-Adesi, University of Lugano, Switzerland

The stability of Factor Models of Interests rates


-  Erhan Cinlar, Princeton University, Princeton, USA

Brownian movement in gamma fields


-  William H. E. Day, Port Maitland , Canada

Biological Aggregation at the Interface Between Theory and Practice

To understand evolutionary processes better, biologists use aggregation methods to estimate evolutionary relationships; yet properties of the methods are sometimes so imprecisely defined, and their interrelationships so poorly understood, that useful formal results may be difficult to obtain. To address this problem I describe a strategy for modeling aggregation methods and studying their properties. Although motivated by biological problems, the strategy and its conceptual framework are generally applicable in research areas where aggregation problems arise.


-  Paul Deheuvels, Laboratoire de Statistique Théorique et Appliquée, Université Paris VI, France

Tests de Cramér-von Mises multivariés


-  Trevor Hastie, Statistics Department, Stanford University, USA

The Entire Regularization Path for the Support Vector Machine

The Support Vector Machine is a widely used tool for classification. Many efficient implementations exist for fitting a two-class SVM model. The user has to supply values for the tuning parameters: the regularization cost parameter, and the kernel parameters. It seems a common practice is to use a default value for the cost parameter, often leading to the least restrictive model. In this paper we argue that the choice of the cost parameter can be critical. We then derive an algorithm that can fit the entire path of SVM solutions for every value of the cost parameter, with essentially the same computational cost as fitting one SVM model. We illustrate our algorithm on some examples, and use our representation to give further insight into the range of SVM solutions.



 

SPIP

Site administration | Home page