m

F Nous contacter


0

Videothèque  | enregistrements trouvés : 10

O
     

-A +A

P Q

Déposez votre fichier ici pour le déplacer vers cet enregistrement.

Multi angle  Spectra of ultra-discrete limits
Zuk, Andrzej (Auteur de la Conférence) | CIRM (Editeur )

We present a computation of spectra of random walks on self-similar graphs.

37A30 ; 05C25 ; 35Q53 ; 20M35

Déposez votre fichier ici pour le déplacer vers cet enregistrement.

It is generally admitted that financial time series have heavy tailed marginal distributions. When time series models are fitted on such data, the non-existence of appropriate moments may invalidate standard statistical tools used for inference. Moreover, the existence of moments can be crucial for risk management. This talk considers testing the existence of moments in the framework of standard and augmented GARCH models. In the case of standard GARCH, even-moment conditions involve moments of the independent innovation process. We propose tests for the existence of moments of the returns process that are based on the joint asymptotic distribution of the estimator of the volatility parameters and empirical moments of the residuals. To achieve efficiency gains we consider non Gaussian QML estimators founded on reparametrizations of the GARCH model, and we discuss optimality issues. We also consider augmented GARCH processes, for which moment conditions are less explicit. We establish the asymptotic distribution of the empirical moment Generating function (MGF) of the model, defined as the MGF of the random autoregressive coefficient in the volatility dynamics, from which a test is deduced. An alternative test is based on the estimation of the maximal exponent characterizing the existence of moments. Our results will be illustrated with Monte Carlo experiments and real financial data.
It is generally admitted that financial time series have heavy tailed marginal distributions. When time series models are fitted on such data, the non-existence of appropriate moments may invalidate standard statistical tools used for inference. Moreover, the existence of moments can be crucial for risk management. This talk considers testing the existence of moments in the framework of standard and augmented GARCH models. In the case of ...

37M10 ; 62M10 ; 62P20

Déposez votre fichier ici pour le déplacer vers cet enregistrement.

Many effects of climate change seem to be reflected not in the mean temperatures, precipitation or other environmental variables, but rather in the frequency and severity of the extreme events in the distributional tails. The most serious climate-related disasters are caused by compound events that result from an unfortunate combination of several variables. Detecting changes in size or frequency of such compound events requires a statistical methodology that efficiently uses the largest observations in the sample.
We propose a simple, non-parametric test that decides whether two multivariate distributions exhibit the same tail behavior. The test is based on the entropy, namely Kullback-Leibler divergence, between exceedances over a high threshold of the two multivariate random vectors. We study the properties of the test and further explore its effectiveness for finite sample sizes.
Our main application is the analysis of daily heavy rainfall times series in France (1976 -2015). Our goal in this application is to detect if multivariate extremal dependence structure in heavy rainfall change according to seasons and regions.
Many effects of climate change seem to be reflected not in the mean temperatures, precipitation or other environmental variables, but rather in the frequency and severity of the extreme events in the distributional tails. The most serious climate-related disasters are caused by compound events that result from an unfortunate combination of several variables. Detecting changes in size or frequency of such compound events requires a statistical ...

Déposez votre fichier ici pour le déplacer vers cet enregistrement.

T.C. Brown and A.R. Freedman proved that the set $\mathcal{P}_{2}$ of products of two primes contains no dense cluster; technically, $\mathcal{P}_{2}$ has a zero upper Banach density, defined as $\delta^{*}(\mathcal{P}_{2}) =\lim_{H\mapsto \infty} \limsup_{x\mapsto \infty} \frac{1}{H} Card \{n\in \mathcal{P}_{2}:x< n\leq x+H\}$.
Pramod Eyyunni, Sanoli Gun and I jointly studied the local behaviour of the product of two shifted primes $\mathcal{Q}_{2}=\{(q-1)(r-1):q,r \, primes\}$. Assuming a classical conjecture of Dickson, we proved that $\delta^{*}(\mathcal{Q}_{2}) = 1/6$. Notice that we know no un-conditional proof that $\delta^{*}(\mathcal{Q}_{2})$ is positive. The application, which was indeed our motivation, concerns the study of the local behaviour of the set $\mathcal{V}$ of values of Euler’s totient function. Assuming Dickson’s conjecture, we prove that $\delta^{*}(\mathcal{V})\geq 1/4$. The converse inequality $\delta^{*}(\mathcal{V})\leq 1/4$ had been proved in the previous millenium by K. Ford, S. Konyagin and C. Pomerance.
T.C. Brown and A.R. Freedman proved that the set $\mathcal{P}_{2}$ of products of two primes contains no dense cluster; technically, $\mathcal{P}_{2}$ has a zero upper Banach density, defined as $\delta^{*}(\mathcal{P}_{2}) =\lim_{H\mapsto \infty} \limsup_{x\mapsto \infty} \frac{1}{H} Card \{n\in \mathcal{P}_{2}:x< n\leq x+H\}$.
Pramod Eyyunni, Sanoli Gun and I jointly studied the local behaviour of the product of two shifted primes $\m...

11B83 ; 11B05 ; 11N32 ; 11N64

Déposez votre fichier ici pour le déplacer vers cet enregistrement.

Discussing recent work joint with M. Rudnev [2], I will discuss the modern approach to the sum-product problem in the reals. Our approach builds upon and simplifies the arguments of Shkredov and Konyagin [1], and in doing so yields a new best result towards the problem. We prove that
$max(\left | A+A \right |,\left | A+A \right |)\geq \left | A \right |^{\frac{4}{3}+\frac{2}{1167}-o^{(1)}}$ , for a finite $A\subset \mathbb{R}$. At the heart of our argument are quantitative forms of the two slogans ‘multiplicative structure of a set gives additive information’, and ‘every set has a multiplicatively structured subset’.
Discussing recent work joint with M. Rudnev [2], I will discuss the modern approach to the sum-product problem in the reals. Our approach builds upon and simplifies the arguments of Shkredov and Konyagin [1], and in doing so yields a new best result towards the problem. We prove that
$max(\left | A+A \right |,\left | A+A \right |)\geq \left | A \right |^{\frac{4}{3}+\frac{2}{1167}-o^{(1)}}$ , for a finite $A\subset \mathbb{R}$. At the heart of ...

11N99 ; 11F99 ; 11B75 ; 11B30 ; 05D10

Déposez votre fichier ici pour le déplacer vers cet enregistrement.

A square in a matrix $\mathcal M =(a_{ij})$ is a 2X2 sub-matrix of $\mathcal M$ with entries $a_{ij}, a_{i+s,j}, ai,j+s, a_{i+s,j+s}$s for some $s\geq 1$. An Erickson matrix is a square binary matrix that contains no squares with constant entries. In [Eri96], Erickson asked for the maximum value of $n$ for which there exists an n x n Erickson matrix. In [AM08] Axenovich and Manske gave an upper bound of around $2^{2^{40}}$. This gargantuan bound was later improved by Bacher and Eliahou in [BE10] using computational means to the optimal value of 15.
In this talk we present the study of a zero-sum analogue of the Erickson matrices problem where we consider binary matrices with entries in {-1,1}. For this purpose, of course, we need to take into account the discrepancy or deviation of the matrix, defined as the sum of all its entries, that is
$disc(\mathcal M)= \sum_{1\leq i\leq n \; \; 1\leq j\leq m}a_{i,j}$.
A zero-sum square is a square $\mathcal S$ with $disc(\mathcal S) = 0$. A natural question is, for example, the following: is it true that for sufficiently large $n$ every $n\times n \{-1,1\} - matrix \, \mathcal M$ with $disc(\mathcal M) = 0$ contains a zero-sum square? We answered positive to this question. Since, our proof uses an induction argument, in order for the induction to work we prove the following stronger statement: For $n \geq 5$ and $m \in \{n,n+1\}$, every $ n \times m \{-1, 1\}$ -matrix $M$ with $\left | disc(M) \right |\leq n$ contains a zero-sum square except for the triangular matrix (up to symmetries), where a triangular matrix is a matrix with all entries above the diagonal equal to -1 and all remaining entries equal to 1.This is a joint work with Edgardo Roldn-Pensado and Alma R. Arvalo.
A square in a matrix $\mathcal M =(a_{ij})$ is a 2X2 sub-matrix of $\mathcal M$ with entries $a_{ij}, a_{i+s,j}, ai,j+s, a_{i+s,j+s}$s for some $s\geq 1$. An Erickson matrix is a square binary matrix that contains no squares with constant entries. In [Eri96], Erickson asked for the maximum value of $n$ for which there exists an n x n Erickson matrix. In [AM08] Axenovich and Manske gave an upper bound of around $2^{2^{40}}$. This gargantuan bound ...

05C55 ; 05D05 ; 11P99

Déposez votre fichier ici pour le déplacer vers cet enregistrement.

The Liouville function $\lambda(n)$ takes the value +1 or -1 depending on whether $n$ has an even or an odd number of prime factors. The Liouville function is closely related to the characteristic function of the primes and is believed to behave more-or-less randomly.
I will discuss my very recent work with Radziwill, Tao, Teräväinen, and Ziegler, where we show that, in almost all intervals of length $X^{\varepsilon}$, the Liouville function does not correlate with polynomial phases or more generally with nilsequences.
I will also discuss applications to superpolynomial number of sign patterns for the Liouville sequence and to a new averaged version of Chowla’s conjecture.
The Liouville function $\lambda(n)$ takes the value +1 or -1 depending on whether $n$ has an even or an odd number of prime factors. The Liouville function is closely related to the characteristic function of the primes and is believed to behave more-or-less randomly.
I will discuss my very recent work with Radziwill, Tao, Teräväinen, and Ziegler, where we show that, in almost all intervals of length $X^{\varepsilon}$, the Liouville function ...

11B30 ; 11N25 ; 11N64

Déposez votre fichier ici pour le déplacer vers cet enregistrement.

In this paper we study asymptotic properties of random forests within the framework of nonlinear time series modeling. While random forests have been successfully applied in various fields, the theoretical justification has not been considered for their use in a time series setting. Under mild conditions, we prove a uniform concentration inequality for regression trees built on nonlinear autoregressive processes and, subsequently, use this result to prove consistency for a large class of random forests. The results are supported by various simulations. (This is joint work with Mikkel Slot Nielsen.)
In this paper we study asymptotic properties of random forests within the framework of nonlinear time series modeling. While random forests have been successfully applied in various fields, the theoretical justification has not been considered for their use in a time series setting. Under mild conditions, we prove a uniform concentration inequality for regression trees built on nonlinear autoregressive processes and, subsequently, use this ...

62G10 ; 60G10 ; 60J05 ; 62M05 ; 62M10

Déposez votre fichier ici pour le déplacer vers cet enregistrement.

In time series analysis there is an apparent dichotomy between time and frequency domain methods. The aim of this paper is to draw connections between frequency and time domain methods. Our focus will be on reconciling the Gaussia likelihood and the Whittle likelihood. We derive an exact, interpretable, bound between the Gaussian and Whittle likelihood of a second order stationary time series. The derivation is based on obtaining the transformation which is biorthogonal to the discrete Fourier transform of the time series. Such a transformation yields a new decomposition for the inverse of a Toeplitz matrix and enables the representation of the Gaussian likelihood within the frequency domain. We show that the difference between the Gaussian and Whittle likelihood is due to the omission of the best linear predictions outside the domain of observation in the periodogram associated with the Whittle likelihood. Based on this result, we obtain an approximation for the difference between the Gaussian and Whittle likelihoods in terms of the best fitting, finite order autoregressive parameters. These approximations are used to define two new frequency domain quasi-likelihoods criteria. We show these new criteria yield a better approximation of the spectral divergence criterion, as compared to both the Gaussian and Whittle likelihoods. In simulations, we show that the proposed estimators have satisfactory finite sample properties.
In time series analysis there is an apparent dichotomy between time and frequency domain methods. The aim of this paper is to draw connections between frequency and time domain methods. Our focus will be on reconciling the Gaussia likelihood and the Whittle likelihood. We derive an exact, interpretable, bound between the Gaussian and Whittle likelihood of a second order stationary time series. The derivation is based on obtaining the tr...

62M10 ; 62M15 ; 62F10

Déposez votre fichier ici pour le déplacer vers cet enregistrement.

The class of integer-valued trawl processes has recently been introduced for modelling univariate and multivariate integer-valued time series with short or long memory.

In this talk, I will discuss recent developments with regards to model estimation, model selection and forecasting of such processes. The new methods will be illustrated in an empirical study of high-frequency financial data.

This is joint work with Mikkel Bennedsen (Aarhus University), Asger Lunde (Aarhus University) and Neil Shephard (Harvard University).
The class of integer-valued trawl processes has recently been introduced for modelling univariate and multivariate integer-valued time series with short or long memory.

In this talk, I will discuss recent developments with regards to model estimation, model selection and forecasting of such processes. The new methods will be illustrated in an empirical study of high-frequency financial data.

This is joint work with Mikkel Bennedsen (Aarhus ...

37M10 ; 60G10 ; 60G55 ; 62F99 ; 62M10 ; 62P05

Z