ENBIS-16 in Sheffield

11 – 15 September 2016; Sheffield Abstract submission: 20 March – 4 July 2016

My abstracts


The following abstracts have been accepted for this event:

  • Choosing Designs for Computer Experiments

    Authors: William Notz (The Ohio State University)
    Primary area of focus / application: Design and analysis of experiments
    Keywords: Designs for computer experiments, Space-filling design, Guidelines, Choosing a design
    Submitted at 29-Jul-2016 12:11 by William Notz
    Accepted (view paper)
    12-Sep-2016 14:30 Choosing Designs for Computer Experiments
    Computer experiments have unique features that require special types of designs. Developing experimental designs for computer experiments, therefore, has become a fruitful research topic. Among the more popular designs are various space-filling designs. However, the justification for using space-filling is sometimes made on intuitive grounds without careful consideration of the goal of the experiment or the statistical model that will be fit to the data. In this talk I will briefly review types of space-filling designs that have been proposed, discuss settings in which these designs are reasonable, and highlight a few results in the literature that provide justification for these designs. I will also propose some guidelines and suggestions for choosing a design for a computer experiment.
  • Privacy, Confidentiality, and Big Data

    Authors: Stephen Fienberg (Department of Statistics, Carnegie Mellon University, Pittsburgh)
    Primary area of focus / application: Other: Big Data
    Secondary area of focus / application: Other:
    Keywords: Privacy, Confidentiality, Big Data, Protection
    Submitted at 2-Aug-2016 10:40 by Jens Bischoff
    As the size and complexity of databases grow, so do concerns regarding the protection of data from inappropriate use by others. On the one hand we want big data to be shareable and useful for a variety of purposes, and on the other hand we know that increased access to more and more data poses increased risk to individual privacy. Is it reasonable to have as our goal maximizing the benefits of big data while minimizing the risk to individual privacy? I review some recent advances in methodology for confidentiality and privacy protection and consider the extent to which they scale to be of use in the sharing of privacy-protected big data sets.
  • EU Funding Session

    Authors: Alessandro Di Bucchianico (Eindhoven University of Technology)
    Primary area of focus / application: Other: Funding
    Keywords: EU funding, Open consultation, COST, ENBIS activities
    Submitted at 2-Aug-2016 12:14 by Alessandro Di Bucchianico
    Accepted (view paper)
    13-Sep-2016 15:40 Special Session: EU Funding Session
    In this session we will update the audience to both past and future actions of ENBIS with respect to European Union funding opportunities.

    The session will start off with a brief overview on the various EU funding opportunities from various points of view. Then we will proceed with a report on the EU Open Consultation on Mathematics in which ENBIS participated actively. Opportunities for ENBIS members will be analysed.
    The final part of the session will be devoted to an update on the status EU COST networking proposal that is being prepared by a group of ENBIS members to support ENBIS activities.
  • Models and Misunderstanding: Statistics, Data Science and Big Data in the Modern World

    Authors: David J. Hand (Imperial College London)
    Primary area of focus / application: Other: George Box Award
    Secondary area of focus / application: Modelling
    Keywords: George Box, Modern data science, Big Data, Statistical model
    Submitted at 9-Aug-2016 15:14 by David Hand
    12-Sep-2016 08:50 George Box Award: David Hand. Award Talk on "Models and Misunderstanding: Statistics, Data Science and Big Data in the Modern World"
    I use George Box’s famous observation that all models are wrong, but some are useful as a framework through which to explore approaches to analysing data. In particular, there are two distinct types of statistical models and, although the distinction is rarely taught in statistics courses, an appreciation of the difference can be useful in statistical modelling and understanding. Different model types are best suited for tackling different kinds of problems. Modern data science, however, and proponents of ‘big data’ in particular, tend to focus on just one of the types, with the consequence that mistakes can be made. Some examples are given.