175 research outputs found
Multiparameter multifractional Brownian motion: Local nondeterminism and joint continuity of the local times
No full textMoving average Multifractional Processes with Random Exponent: lower bounds for local oscillations
No full textInternational audienceIn the last few years Ayache, Esser and Hamonier introduced a new Multifractional Process with Random Exponent (MPRE) obtained by replacing the Hurst parameter in a moving average representation of Fractional Brownian Motion through Wiener integral by an adapted Hölder continuous stochastic process indexed by the integration variable. Thus, this MPRE can be expressed as a moving average Itô integral which is a considerable advantage with respect to another MPRE introduced a long time ago by Ayache and Taqqu. Thanks to this advantage, very recently, Loboda, Mies and Steland have derived interesting results on local Hölder regularity, self-similarity and other properties of the recently introduced moving average MPRE and generalizations of it. Yet, the problem of obtaining, on an universal event of probability 1 not depending on the location, relevant lower bounds for local oscillations of such processes has remained open. We solve it in the present article under some conditions
My Matter
No full textAn interview with the writer, financial technologist and author of Blank Swan, Elie Ayache by artist Roman Vasseur on the ramifications for the visual arts of Ayache’s work on probability, contingency and matter in the financial markets
Bourgeois du faubourg Saint-Antoine 1791 – 1792
No full textAfin de réaliser cette étude, j’ai exploité un échantillonnage strictement délimité de 71 prosopographies sociales et politiques issues de deux mémoires de maîtrise, celui d’I. Ayache, portant sur les électeurs du faubourg en 1792, et le mien, concernant les électeurs de 1791. Essayons d’identifier ces bourgeois du faubourg Saint-Antoine qui sont nouvellement élus au suffrage censitaire sous la Constituante et la Législative. Quelle place occupent-ils économiquement au sein du faubourg, et qu..
Linear fractional stable sheets: Wavelet expansion and sample path properties
No full textAbstractIn this paper we give a detailed description of the random wavelet series representation of real-valued linear fractional stable sheet introduced in [A. Ayache, F. Roueff, Y. Xiao, Local and asymptotic properties of linear fractional stable sheets, C.R. Acad. Sci. Paris Ser. I. 344 (6) (2007) 389–394]. By using this representation, in the case where the sample paths are continuous, an anisotropic uniform and quasi-optimal modulus of continuity of these paths is obtained as well as an upper bound for their behavior at infinity and around the coordinate axes. The Hausdorff dimensions of the range and graph of these stable random fields are then derived
Linear fractional stable sheets: Wavelet expansion and sample path properties
No full textIn this paper we give a detailed description of the random wavelet series representation of real-valued linear fractional stable sheet introduced in [A. Ayache, F. Roueff, Y. Xiao, Local and asymptotic properties of linear fractional stable sheets, C.R. Acad. Sci. Paris Ser. I. 344 (6) (2007) 389-394]. By using this representation, in the case where the sample paths are continuous, an anisotropic uniform and quasi-optimal modulus of continuity of these paths is obtained as well as an upper bound for their behavior at infinity and around the coordinate axes. The Hausdorff dimensions of the range and graph of these stable random fields are then derived.Wavelet analysis Stable processes Linear fractional stable sheet Modulus of continuity Hausdorff dimension
On filter-type estimation of discretely sampled cyclic long-memory processes
No full textThe generalized filtered method of moments was developed in the recent papers by Alomari et al., 2020, and Ayache et al., 2022. It used functional data obtained from continuously sampled cyclic long-memory stochastic processes to simultaneously estimate their parameters. However, the majority of applications deal with discretely sampled processes or time series. This paper extends the approach to accommodate discrete-time scenarios. It proves that the new discrete estimates exhibit analogous properties to the continuous case and are strongly consistent with the same rates of convergence. The numerical study results are presented to illustrate the theoretical findings and to indicate the sampling rates and resolution levels required for accurate estimates.22 pages, 6 figure
Recent Developments in Fractals and Related Fields
No full text18 pagesIn Ayache and Taqqu (2005), the multifractional Brownian (mBm) motion is obtained by replacing the constant parameter of the fractional Brownian motion (fBm) by a smooth enough functional parameter depending on the time . Here, we consider the process obtained by replacing in the wavelet expansion of the fBm the index by a function depending on the dyadic point . This process was introduced in Benassi et al (2000) to model fBm with piece-wise constant Hurst index and continuous paths. In this work, we investigate the case where the functional parameter satisfies an uniform Hölder condition of order \beta>\sup_{t\in \rit} H(t) and ones shows that, in this case, the process is very similar to the mBm in the following senses: i) the difference between and a mBm satisfies an uniform Hölder condition of order ; ii) as a by product, one deduces that at each point the pointwise Hölder exponent of is and that is tangent to a fBm with Hurst parameter
Statistical inference for hidden multifractionnal processes in a setting of stochastic volatility models
No full textL’exemple paradigmatique d’un processus stochastique multifractionnaire est le mouvement brownien multifractionnaire (mbm). Ce processus gaussien de nature fractale admet des trajectoires continues nulle part dérivables et étend de façon naturelle le célèbre mouvement brownien fractionnaire (mbf). Le mbf a été introduit depuis longtemps par Kolmogorov et il a ensuite été « popularisé » par Mandelbrot ; dans plusieurs travaux remarquables, ce dernier auteur a notamment insisté sur la grande importance de ce modèle dans divers domaines applicatifs. Le mbm, quant à lui, a été introduit, depuis plus de quinze ans, par Benassi, Jaffard, Lévy Véhel, Peltier et Roux. Grossièrement parlant, il est obtenu en remplaçant le paramètre constant de Hurst du mbf, par une fonction H(t) qui dépend de façon régulière du temps t. Ainsi, contrairement au mbf, les accroissements du mbm sont non stationnaires et la rugosité locale de ses trajectoires (mesurée habituellement par l’exposant de Hölder ponctuel) peut évoluer significativement au cours du temps ; en fait, à chaque instant t, l’exposant de Hölder ponctuel du mbm vaut H(t). Notons quecette dernière propriété, rend ce processus plus flexible que le mbf ; grâce à elle, le mbm est maintenant devenu un modèle utile en traitement du signal et de l’image ainsi que dans d’autres domaines tels que la finance. Depuis plus d’une décennie, plusieurs auteurs se sont intéressés à des problèmes d’inférence statistique liés au mbm et à d’autres processus/champs multifractionnaires ; leurs motivations comportent à la fois des aspects applicatifs et théoriques. Parmi les plus importants, figure le problème de l’estimation de H(t), l’exposant de Hölder ponctuel en un instant arbitraire t. Dans ce type de problématique, la méthode des variations quadratiques généralisées, initialement introduite par Istas et Lang dans un cadre de processus à accroissements stationnaires, joue souvent un rôle crucial. Cette méthode permet de construire des estimateurs asymptotiquement normaux à partir de moyennes quadratiques d’accroissements généralisés d’un processus observé sur une grille. A notre connaissance, dans la littérature statistique qui concerne le mbm, jusqu’à présent, il a été supposé que, l’observation sur une grille des valeurs exactes de ce processus est disponible ; cependant une telle hypothèse ne semble pas toujours réaliste. L’objectif principal de la thèse est d’étudierdes problèmes d’inférence statistique liés au mbm, lorsque seulement une version corrompue de ce dernier est observable sur une grille régulière.Cette version corrompue est donnée par une classe de modèles à volatilité stochastique dont la définition s’inspire de certains travaux antérieurs de Gloter et Hoffmann ; signalons enfin que la formule d’Itô permet de ramener ce cadre statistique au cadre classique : « signal+bruit ».The paradigmatic example of a multifractional stochastic process is multifractional Brownian motion (mBm). This fractal Gaussian process with continuous nowhere differentiable trajectories is a natural extension of the well-known fractional Brownian motion (fBm). FBm was introduced a longtime ago by Kolmogorov and later it has been made « popular» by Mandelbrot; in several outstanding works, the latter author has emphasized the fact that this model is of a great importance in various applied areas. Regarding mBm, it was introduced, more than fifteen years ago, by Benassi, Jaffard, Lévy Véhel, Peltier and Roux. Roughly speaking, it is obtained by replacing the constant Hurst parameter of fBm by a smooth function H(t) which depends on the time variable t. Therefore, in contrast with fBm, theincrements of mBm are non stationary and the local roughness of its trajectories (usually measured through the pointwise Hölder exponent) is allowed to significantly evolve over time; in fact, at each time t, the pointwise Hölder exponent of mBm is equal to H(t). It is worth noticing that the latter property makes this process more flexible than fBm; thanks to it, mBm has now become a useful model in the area of signal and image processing, aswell as in other areas such as finance. Since at least one decade, several authors have been interested in statistical inference problems connected with mBm and other multifractional processes/fields; their motivations have both applied and theoretical aspects. Among those problems, an important one is the estimation of H(t), the pointwise Hölder exponent at an arbitrary time t. In the solutions of such issues, the generalized quadratic variation method, which was first introduced by Istas and Lang in a setting of stationary increments processes, usually plays a crucial role. This method allows to construct asymptotically normal estimators starting from quadratic means of generalized increments of a process observed on a grid. So far, to our knowledge, in the statistical literature concerning mBm, it has been assumed that, the observation of the true values of this process on a grid, is available; yet, such an assumption does not always seem to be realistic. The main goal of the thesis is to study statistical inference problems related to mBm, when only a corrupted version of it, can be observed on a regular grid. This corrupted version is given by a class of stochastic volatility models whose definition is inspired by some Gloter and Hoffmann’s earlier works; last, notice that thanks to Itô formula this statistical setting can be viewed as the classical setting: « signal+noise »
Wavelet analysis of stationary increments harmonizable stable fields
No full textL’étude du comportement trajectoriel des champs/processus stochastiques est un sujet de recherche classique en théorie des probabilités et dans des domaines connexes comme la géométrie fractale. Dans cet objectif, plusieurs méthodes ont été développées depuis longtemps afin d’étudier le comportement des trajectoires de champs/processus gaussiens. Ces méthodes reposent souvent sur une structure hilbertienne « sympathique », et peuvent aussi nécessiter la finitude de moments d’ordre élevé. Ainsi, elles sont difficilement transposables dans des cadres de lois à queue lourde. Ces dernières sont importantes en probabilités et en statistique parce qu’elles constituent une contrepartie naturelle des lois gaussiennes. Dans le cas de certains champs/processus stables linéaires de type moyenne mobile non anticipative, tels que le drap fractionnaire stable linéaire et le mouvement multifractionnaire stable linéaire, des méthodes d’ondelettes, assez nouvelles, se sont déjà avérées fructueuses dans l’étude du comportement trajectoriel. Peut-on adapter cette méthodologie à certains champs/processus stables harmonisables ? Donner une réponse à cette question est un problème assez délicat car, de façon générale, de grandes différences séparent le cadre stable harmonisable de celui de type moyenne mobile. Le principal objectif de la thèse est d’étudier cette question dans le cadre d’un champ stable harmonisable symétrique à accroissement stationnaire de forme générale.Studying sample path behaviour of stochastic fields/processes is a classical research topic in probability theory and related areas such as fractal geometry. To this end, many methods have been developed for a long time in order to study sample path behaviour of Gaussian fields/processes. They often rely on some underlying "nice" Hilbertian structure, and can also require finiteness of moments of high order. Therefore, they can hardly be transposed to frames of heavy-tailed stable probability distributions. Such distributions are very important in probability and statistics because they are a natural counterpart to the Gaussian ones. In the case of some linear non-anticipative moving average stable fields/processes, such as the linear fractional stable sheet and the linear multifractional stable motion, rather new wavelet methods have already proved to be successful in studying sample path behaviour. Can this methodology be adapted to some harmonizable stable fields/processes? Providing an answer to this question is a non trivial problem, since, generally speaking, there are large differences between an harmonizable stable setting and a moving average one. The main goal of the thesis is to study this issue in the case of a stationary increments symmetric stable harmonizable field of a general form
- …
