Information détaillée concernant le cours

[ Retour ]
Titre

Ecole d’été 2017

Dates

3-6 septembre 2017

Responsable de l'activité

Yves Tillé

Organisateur(s)/trice(s)

M. Yves Tillé, UNINE (Président), Mme Caroline Gillardin, UNINE (coordinatrice)

Intervenant-e-s

Prof. Sudipto Banerjee, UCLA Prof. Andrew Harvey, University of Cambridge Prof. Jonathan Taylor, Stanford University (invité par l'EPFL)

Description

 

 

Professor Sudipto Banerjee

(UCLA)

Title : High-dimensional Bayesian Geostatistics

 

Abstract

With the growing capabilities of Geographic Information Systems (GIS) and user-friendly software, statisticians today routinely encounter geographically referenced data containing observations from a large number of spatial locations and time points. Over the last decade, hierarchical spatiotemporal process models have become widely deployed statistical tools for researchers to better understand the complex nature of spatial and temporal variability. However, fitting hierarchical spatiotemporal models often involves expensive matrix computations with complexity increasing in cubic order for the number of spatial locations and temporal points. This renders such models unfeasible for large data sets. This article offers a focused review of two methods for constructing well-defined highly scalable spatiotemporal stochastic processes. Both these processes can be used as "priors" for spatiotemporal random fields. The first approach constructs a low-rank process operating on a lower-dimensional subspace. The second approach constructs a Nearest-Neighbor Gaussian Process (NNGP) that ensures sparse precision matrices for its finite realizations. Both processes can be exploited as a scalable prior embedded within a rich hierarchical modeling framework to deliver full Bayesian inference. These approaches can be described as model-based solutions for big spatiotemporal datasets. The models ensure that the algorithmic complexity has $\sim n$ floating point operations (flops), where $n$ the number of spatial locations (per iteration). We compare these methods and provide some insight into their methodological underpinnings.

PDF Lectures1a

PDF Lectures 1b

PDF Lectures 2

PDF Lectures 3

Professor Andrew Harvey

(University of Cambridge)

Title : Modelling Volatility

 

Abstact

1 Nonlinear models and changing volatilityUnobserved components. Introduction to DCS/GAS models. Nonlinear models: independence, uncorrelatedness and martingale di¤erences. Distributions and heavy tails. Properties of nancial returns. Standard volatility models: GARCH, EGARCH and stochastic volatility. Intra-day data, realized volatility, range and duration.
2 Dynamic conditional score (DCS) models
Location and robustness. EGARCH. Leverage, long memory, components. Models for positive variables. Multivariate GARCH models. Dynamic correlation.
3 Recent developments in DCS modelling
Dynamic copulas. Multivariate covariance matrices. Changing shape. Quantiles. Censoring and zeroes. Adaptive ltering. Main references: Harvey, A. C. (2013) Dynamic Models for Volatility and Heavy Tails. Cambridge University Press. Creal, D., Koopman, S.J., and A. Lucas (2013). Generalized autoregres- sive score models with applications. Journal of Applied Econometrics, 28, 777-95. Creal, D., Koopman, S.J. and A. Lucas (2011). A Dynamic Multivariate Heavy-Tailed Model for Time-Varying Volatilities and Correlations, Journal of Business and Economic Statistics, 29, 552-63. Websites econ.cam.ac.uk/DCS; gasmodel.com

PDF Lectures 1

PDF Lectures 2

PDF Lectures 3

Professor Jonathan Taylor

(Stanford University)

Title : Critical points of random functions : geometry and inference

 

Abstact

The topic of the lectures will broadly be about the properties of critical points of random functions with particular emphasis on statistical inference. We will begin with a brief introduction to smooth Gaussian processes and their path properties, with an application to signal detection with smooth background noise. One of the main tools used in the calculations is the Kac-Rice formula which is used to compute the certain number of critical points of a given type. Broadly speaking, the Kac-Rice formula serves as a model for conditioning a smooth function to have a critical point in a given location. This notion of conditioning is important for the final part of the lectures. In the final part of the lectures, our random functions are objective functions of (random) convex programs, we discuss selective inference in which data analysts are allowed to observe some aspects of the solution to the convex program, i.e. some aspects of a given critical point. Particular importance is placed on the concrete example of the LASSO, specifically inference after selection through the LASSO.

Lecture 1

Lecture 2

Lecture 3

 

 

 Programme

 

 

Sunday 03.09

Monday 04.09

Tuesday 05.09

Wednesday 06.09

 8h30-10h00

 

Andrew Harvey

Andrew Harvey

Andrew Harvey

10h00-10h30

 

Coffee Break

Coffee Break

Coffee Break

10h30-12h00

 

Jonathan Taylor

Jonathan Taylor

Jonathan Taylor

12h00-14h00

 

 

 

 

14h00-15h00

 

 

 

 

15h30-17h00

Welcome tea

Coffee Break

Coffee Break

 

17h00-18h30

Sudipto Banerjee

Sudipto Banerjee

Sudipto Banerjee

18h30-19h30

Apero

Apero

Commission

19h30-21h00

Dinner

Dinner

Dinner

 

Lieu

Eurotel Victoria VILLARS

Information

 

 

 Eurotel Victoria Villars
Route des Layeux, 1884 Villars-sur-Ollon (Tél +41 24 495 31 31)

Site Eurotel : http://www.eurotel-victoria.ch/villars/fr Accès à Villars :

 

  • EN VOITURE : Autoroute A9, direction Grand St-Bernard, sortie Aigle. Puis suivre les panneaux routiers « Ollon » puis « Villars ».
  • EN AVION : Aéroports internationaux de:

- Genève (120 km) ou Zürich (250 km) ou Bâle (200 km)
- puis EN TRAIN  (HORAIRE DES TRAINS - RAILWAY TIMETABLE)

  • EN TRAIN ET BUS : Suisse Train schedule : From: Geneve airport, To: Villars-sur-Ollon, gare. Trains directs jusqu'à Aigle. Ensuite bus 14429 (Aigle - Villars-sur-Ollon gare).
  • EN TRAIN : Suisse Train schedule : From: Geneve airport, To: Bex, gare. Trains directs jusqu'à Bex. Changer de train à Bex jusqu'à Villars.


Durée des trajets:
Lausanne - Aigle (30 minutes), Aigle - Villars-sur-Ollon (35 minutes). Visa pour la Suisse (Swiss Online Visa application) Météo en suisse (meteoswiss.admin.ch)

Frais

Doctorant CUSO chambre double: 200 CHF
Doctorant CUSO chambre simple: 350 CHF
Post-doctorant CUSO chambre double: 300 CHF
Post-doctorant CUSO chambre simple: 450 CHF
Professeur CUSO chambre double: 400 CHF
Professeur CUSO chambre simple: 550 CHF
Non CUSO universitaire chambre double: 850 CHF
Non CUSO universitaire chambre simple: 1000
Non CUSO privé chambre double: 1300 CHF
Non CUSO privé chambre simple: 1500 CHF

Lors de votre inscription, merci de bien vouloir indiquer dans la zone commentaire si vous désirez une chambre simple, ou double et le nom de la personne avec qui vous souhaiteriez partager votre chambre. Dans le cas où rien n'est indiqué, une chambre simple sera réservée.

Inscription

Versement sur compte postal:

CUSO
CCP 12-1873-8
Neuchâtel
BIC : POFICHBEXXX
IBAN : CH0509000000120018738 Merci d'écrire votre nom lors du paiement

Places

32

Délai d'inscription 01.09.2017
short-url short URL

short-url URL onepage