4th International Conference of the ERCIM WG on
COMPUTING & STATISTICS (ERCIM'11)
17-19 December 2011, Senate House, University of London, UK

Home Call For Papers Committees Plenary Talks Tracks &
Organized Sessions Invited Sessions Tutorials Programme Summary Programme Abstract Submission Registration Special Issues Venue, Travel & Accommodation Sponsors Working Group Past Conferences Contact


TUTORIALS


The tutorials will take place on Friday the 16th of December 2011 at the Senate House. The number of participants to the tutorials is limited and restricted only to those who attend the conference. For further information please contact Ana Colubi.



Tutorial Programme
Friday 16th December 2011


Tutorial 1 (by Esther Ruiz): Modelling conditional means and variances: differences between macro and financial time series
Room: Senate Room at the Senate House

  The lecture notes which will be used during the tutorial can be obtained from the file Time_Series.pdf  

09:00 - 10:35   Lecture
10:30 - 11:00   Coffee Break (Served at the top of the ceremonial stairs, Senate House)
11:00 - 12:00   Lecture
12:00 - 12:15   Pause
12:15 - 13:30   Lecture

Tutorial 2 (by Christian Hennig): The Gaussian mixture model as a tool for cluster analysis
Room: Jessel Room at the Senate House

15:00 - 17:00   Lecture
17:00 - 17:30   Coffee Break (Served at the top of the ceremonial stairs, Senate House)
17:30 - 19:30   Lecture

TUTORIAL 1: Modelling conditional means and variances: differences between macro and financial time series

Room: Senate Room at the Senate House      Time: 9:00 - 13:30

Esther Ruiz, Department of Statistics, Universidad Carlos III de Madrid, Spain. Email: Contact

  The lecture notes which will be used during the tutorial can be obtained from the file Time_Series.pdf  

Economic and financial time series, as any other time series, share several properties and therefore have to be represented by stochastic processes. However, when dealing with macroeconomic variables, most of the interest focus on modelling the conditional means with conditional variances having a secondary interest. On the other hand, in the context of financial time series, the conditional means are usually rather constant and the focus is on conditional variances. Although the statistical fundamentals in time series analysis is the same regardless of whether we look at macro or financial time series, in this tutorial, we will revise the main models implemented in each case. In this context, topics as nonlinear relationships, white noise versus strict white noise, conditional means and variances versus the corresponding marginal moments are cruzial for an adequate analysis of the data. The empirical examples will be illustrated using Eviews-7.

TUTORIAL 2: The Gaussian mixture model as a tool for cluster analysis

Room: Jessel Room at the Senate House      Time: 15:00 - 19:30

Christian Hennig, Department of Statistical Science, University College London , UK. Email: Contact

The Gaussian mixture model is one of the most versatile tools for cluster analysis, based on modelling every cluster by a multivariate Gaussian distribution. In this tutorial, the Gaussian mixture model will be introduced, along with the use of the R-package mclust for estimation and classification of the points. Example datasets will be analysed. Theory will be presented as far as necessary to understand the results in practice.


The tutorial will also treat methods to validate the resulting clustering by exploring robustness, stability and separateness of the clusters. Shortcomings of the methodology will be discussed, as well as criteria to decide whether the Gaussian mixture model delivers a suitable clustering in a given situation, or whether rather different clustering methods should be used. All material will be illustrated by example datasets and computer code.



Lunch possibilities


There are various sandwich places and Cafes around the Senate House. There is a Costa Cafe in the ground level of the building and another one at the Birkbeck University main building. The Senate House is 10 minutes walk from Russell Square and other places in Bloomsbury where there is a choice of restaurants. (Please check the Area Map).