Search

Despite the widespread adoption of Large language models (LLMs), their remarkable capabilities remain limited to a few high-resource languages. Additionally, many low-resource languages (e.g. African languages) are often evaluated only on basic text classification tasks due to the lack of appropriate or comprehensive benchmarks outside of high-resource languages. In this paper, we introduce IrokoBench -- a human-translated benchmark dataset for 16 typologically-diverse low-resource African languages covering three tasks: natural language inference~(AfriXNLI), mathematical reasoning~(AfriMGSM), and multi-choice knowledge-based QA~(AfriMMLU). We use IrokoBench to evaluate zero-shot, few-shot, and translate-test settings~(where test sets are translated into English) across 10 open and four proprietary LLMs. Our evaluation reveals a significant performance gap between high-resource languages~(such as English and French) and low-resource African languages. We observe a significant performance gap between open and proprietary models, with the highest performing open model, Aya-101 only at 58\% of the best-performing proprietary model GPT-4o performance. Machine translating the test set to English before evaluation helped to close the gap for larger models that are English-centric, like LLaMa 3 70B. These findings suggest that more efforts are needed to develop and adapt LLMs for African languages., Comment: Under review

In this paper, we study some aspects on random analysis on the L\'eevy stochastic processes with margins following generalized hyperbolic distributions generated by gamma laws. In particular we study the boundedness of its total variations and the quadratic variations. Next we give an empirical construction that enables the graphical representation of the paths of such stochastic processes. Comparisons with the Brownian motions are considered., Comment: 21 pages, 4 figures, Construction of the gamma-gh L\'evy process for

According to the Chinese Health Statistics Yearbook, in 2005, the number of traffic accidents was 187781 with total direct property losses of 103691.7 (10000 Yuan). This research aims to fill the gap in the literature by investigating the extreme claim sizes not only for the entire portfolio. This empirical study investigates the behavior of the upper tail of the claim size by class of policyholders.

We propose a probabilistic setting in which we study the probability law of the Rajaraman and Ullman \textit{RU} algorithm and a modified version of it denoted by \textit{RUM}. These algorithms aim at estimating the similarity index between huge texts in the context of the web. We give a foundation of this method by showing, in the ideal case of carefully chosen probability laws, the exact similarity is the mathematical expectation of the random similarity provided by the algorithm. Some extensions are given. \noindent \textbf{R\'{e}sum\'{e}.} Nous proposons un cadre probabilistique dans lequel nous \'{e}tudions la loi de probabilit\'{e} de l'algorithme de Rajaraman et Ullman \textit{RU} ainsi qu'une version modifi\'{e}e de cet algorithme not\'{e}e \textit{RUM}. Ces alogrithmes visent \`{a} estimer l'indice de la similarit\'{e} entre des textes de grandes tailles dans le contexte du Web. Nous donnons une base de validit\'e de cette m\'{e}thode en montrant que pour des lois de probabilit\'{e}s minutieusement choisies, la similarit\'{e} exacte est l'esp\'{e}rance math\'{e}matique de la similarit\'{e} al\'{e}atoire donn\'{e}e par l'algorithme \textit{RUM}. Des g\'en\'eralisations sont abord\'ees.

The investigation asymptotic limits on associated data mainly focused on limit theorems of summands of associated data and on the related invariance principles. In a series of papers, we are going to set the general frame of the theory by considering an arbitrary infinitely decomposable (divisible) limit law for summands and study the associated functional laws converging to L\'evy processes. The asymptotic frame of Newman (1980) is still used as a main tool. Detailed results are given when $G$ is a Gaussian law (as confirmation of known results) and when $G$ is a Poisson law. In the later case, classical results for independent and identically distributed data are extended to stationary and non-stationary associated data.

The simple L\'evy Poisson process and scaled forms are explicitly constructed from partial sums of independent and identically distributed random variables and from sums of non-stationary independent random variables. For the latter, the weak limits are scaled Poisson processes. The method proposed here prepares generalizations to dependent data, to associated data in the first place., Comment: 21 pages

In earlier stages in the introduction to asymptotic methods in probability theory, the weak convergence of sequences $(X_n)_{n\geq 1}$ of Binomial of random variables (\textit{rv}'s) to a Poisson law is classical and easy-to prove. A version of such a result concerning sequences $(Y_n)_{n\geq 1}$ of negative binomial \textit{rv}'s also exists. In both cases, $X_n$ and $Y_n-n$ are by-row sums $S_n[X]$ and $S_n[Y]$ of arrays of Bernoulli \textit{rv}'s and corrected geometric \textit{rv}'s respectively. When considered in the general frame of asymptotic theorems of by-row sums of \textit{rv}'s of arrays, these two simple results in the independent and identically distributed scheme can be generalized to non-stationary data and beyond to non-stationary and dependent data. Further generalizations give interesting results that would not be found by direct methods. In this paper, we focus on generalizations to the non-stationary independent data. Extensions to dependent data will addressed later., Comment: 21 pages, 0 figure

We introduce a new generalization of the Pseudo-Lindley distribution by applying alpha power transformation. The obtained distribution is referred as the Pseudo-Lindley alpha power transformed distribution (\textit{PL-APT}). Some tractable mathematical properties of the \textit{PL-APT} distribution as reliability, hazard rate, order statistics and entropies are provided. The maximum likelihood method is used to obtain the parameters' estimation of the \textit{PL-APT} distribution. The asymptotic properties of the proposed distribution are discussed. Also, a simulation study is performed to compare the modeling capability and flexibility of \textit{PL-APT} with Lindley and Pseudo-Lindley distributions. The \textit{PL-APT} provides a good fit as the Lindley and the Pseudo-Lindley distribution. The extremal domain of attraction of \textit{PL-APT} is found and its quantile and extremal quantile functions studied. Finally, the extremal value index is estimated by the double-indexed Hill's estimator (Ngom and Lo, 2016) and related asymptotic statistical tests are provided and characterized.

For many probability laws, in parametric models, the estimation of the parameters can be done in the frame of the maximum likelihood method, or in the frame of moment estimation methods, or by using the plug-in method, etc. Usually, for estimating more than one parameter, the same frame is used. We focus on the moment estimation method in this paper. We use the instrumental tool of the functional empirical process (fep) in Lo (2016) to show how it is practical to derive, almost algebraically, the joint distribution Gaussian law and to derive omnibus chi-square asymptotic laws from it. We choose four distributions to illustrate the method (Gamma law, beta law, Uniform law and Fisher law) and completely describe the asymptotic laws of the moment estimators whenever possible. Simulations studies are performed to investigate for each case the smallest sizes for which the obtained statistical tests are recommendable. Generally, the omnibus chi-square test proposed here work fine with sample sizes around fifty, Comment: 36 pages; six figures

This paper investigates the probability density function ($pdf$) of the $(2n-1)$-vector $(n\geq 1)$ of both lower and upper record values for a sequence of independent random variables with common $pdf f$ defined on the same probability space, provided that the lower and upper record times are finite up to $n$. A lot is known about the lower or the upper record values when they are studied separately. When put together, the challenges are a far bigger complicated. The rare results in the literature still present important flaws. This paper begins a new and complete investigation with a few number of records: 2 and 3. Lessons from these simple cases will allow addressing the general formulation of simultaneous joint lower-upper records., Comment: 63 pages, 3 figures

General Central limit theorem deals with weak limits (in type) of sums of row-elements of array random variables. In some situations as in the invariance principle problem, the sums may include only parts of the row-elements. For strictly stationary arrays (stationary for each row), there is no change to the asymptotic results. But for non-stationary data, especially for dependent data, asymptotic laws of partial sums moving in rows may require extra-conditions to exist. This paper deals with central limit theorems with Gaussian limits for non-stationary data. Our main focus is on dependent data, particularly on associated data. But the non-stationary independent data is also studied as a learning process. The results are applied to finite-distributional invariance principles for the types of data described above. In Moreover, results for associated sequences are interesting and innovative. Beyond their own interest, the results are expected to be applied for random sums of random variables and next in statistical modeling in many disciplines, in Actuarial sciences for example, Comment: 43 pages

A uniformly continuously integrable sequence of real-valued measurable functions, defined on some probability space, is relatively compact in the $\sigma(L^1,L^\infty)$ topology. In this paper, we link such a result to weak convergence theory of bounded measures as exposed in Billingsley (1968) and in Lo(2021) to offer a detailed and new proof using the ideas beneath the proof of prohorov's theorem where the continuous integrability replaces the uniform or asymptotic tightness.

This paper is part of series on self-contained papers in which a large part, if not the full extent, of the asymptotic limit theory of summands of independent random variables is exposed. Each paper of the series may be taken as review exposition but specially as a complete exposition expect a few exterior resources. For graduate students and for researchers (beginners or advanced), any paper of the series should be considered as a basis for constructing new results. The contents are taken from advanced books but the organization and the proofs use more recent tools, are given in more details and do not systematically follow previous one. Sometimes, theorems are completed and innovated, Comment: 8 pages

In this paper we investigate the Burr distributions family which contains twelve members. Second order expansions of quantiles of the Burr's distributions are provided on which may be based statistical methods, in particular in extreme value theory. Beyond the proper interest of these expansions, we apply them to characterize the asymptotic laws of their records of Burr's distributions, lead to new statistical tests.

With the success of large-scale pre-training and multilingual modeling in Natural Language Processing (NLP), recent years have seen a proliferation of large, web-mined text datasets covering hundreds of languages. We manually audit the quality of 205 language-specific corpora released with five major public datasets (CCAligned, ParaCrawl, WikiMatrix, OSCAR, mC4). Lower-resource corpora have systematic issues: At least 15 corpora have no usable text, and a significant fraction contains less than 50% sentences of acceptable quality. In addition, many are mislabeled or use nonstandard/ambiguous language codes. We demonstrate that these issues are easy to detect even for non-proficient speakers, and supplement the human audit with automatic analyses. Finally, we recommend techniques to evaluate and improve multilingual corpora and discuss potential risks that come with low-quality data releases., Comment: Accepted at TACL; pre-MIT Press publication version

As the first paper of a series of exploratory analysis and statistical investigation works on the Gambian \textit{GABECE} data based on a variety of statistical tools, we wish to begin with a thorough unsupervised learning process through descriptive and exploratory methods. This will lead to a variety of discoveries and hypotheses that will direct future research works related to this data., Comment: 23 pages; 5 figures

In this paper we want to find a statistical rule that assigns a passing or failing grade to students who undertook at least three exams out of four in a national exam, instead of completely dismissing them students. While it is cruel to declare them as failing, especially if the reason for their absence it not intentional, they should have demonstrated enough merit in the three exams taken to deserve a chance to be declared passing. We use a special classification method and nearest neighbors methods based on the average grade and on the most modal grade to build a statistical rule in a supervised learning process. The study is built on the national GABECE educational data which is a considerable data covering seven years and all the six regions of the Gambia., Comment: 34 pages

In this paper we extend results obtained in [3] and [5]. By considering a semi linear conservation law with velocity in $L^\infty$, we prove by fixed point arguments existence and uniqueness result and even in a penalized situation.

In this note, we combine the two approaches of Billingsley (1998) and Cs\H{o}rg\H{o} and R\'ev\'esz (1980), to provide a detailed sequential and descriptive for creating s standard Brownian motion, from a Brownian motion whose time space is the class of non-negative dyadic numbers. By adding the proof of Etemadi's inequality to text, it becomes self-readable and serves as an independent source for researches and professors., Comment: 17 pages

The moment problem is an important problem in Functional Analysis and in Probability measure. It goes back to Stieltjes, around 1890. There is still an important ongoing interest in the recent literature. But, up today, the main theoretical resource (Shohat and Tamarkin, 1934) does not have the modern exposure it deserves, especially in the current development of measure theory of integration. Besides, the multivariate version is far less exploited. In this paper, a full exposure of such a theory is presented, using the latest knowledge of measure theory and functional analysis. As a result, the basis of future development is layed out and the accessibility of the theory by modern graduate students andresearches is guaranteed., Comment: 28 pages

The representation Skorohod theorem of weak convergence of random variables on a metric space goes back to Skorohod (1956) in the case where the metric space is the class of real-valued functions defined on [0,1] which are right-continuous and have left-hand limits when endowed with the Skorohod metric. Among the extensions of that to metric spaces, the version by Wichura (1970) seems to be the most fundamental. But the proof of Wichura seems to be destined to a very restricted public. We propose a more detailed proof to make it more accessible at the graduate level. However we do far more by simplifying it since important steps in the original proof are dropped, which leads to a direct proof that we hope to be more understandable to a larger spectrum of readers. The current version is more appropriate for different kinds of generalizations.

Integrating with respect to functions which are constant on intervals whose bounds are discontinuity points (of those functions) is frequent in many branches of Mathematics, specially in stochastic processes. For such functions and alike extension, a comparison between Riemann-Stieltjes and Lebesgue-Stieltjes integration and the integrals formulas leads to interesting facts for students (as complements of Measure Theory and Integrations) and for practitioners and and researchers. We undergone conditions of existence the Riemann-Stieltjes integrals related to that type of function and compare the results with what should be expected for Lebesgue-Stieltjes theory., Comment: 10

A method of estimating the joint probability mass function of a pair of discrete random variables is described. This estimator is used to construct the conditional Shannon-R\'eyni-Tsallis entropies estimates. From there almost sure rates of convergence and asymptotic normality are established. The theorical results are validated by simulations., Comment: 28 pages, 4 figures

The pseudo-Lindley distribution which was introduced in Zeghdoudi and Nedjar (2016) is studied with regards to its upper tail. In that regard, and when the underlying distribution function follows the Pseudo-Lindley law, we investigate the behavior of its values, the asymptotic normality of the Hill estimator and the double-indexed generalized Hill statistic process (Ngom and Lo), the asymptotic normality of the records values and the moment problem.

The pseudo-Lindley distribution was introduced as a useful generalization of the Lindley distribution in Zeghdoudi and Nedjar (2016) who showed interesting properties of their new laws and efficiencies in modeling data in Reliability and Survival Analysis. In this paper we study the estimators of the pair of parameters and determine their asymptotic law from which a chi-square law is derived. From both asymptotic laws, statistical tests are built. Simulation studies on the tests conclude to their efficiency for data sizes generally used in Reliability. R codes related to statistical analysis on that law are given in an appropriate archive repository code paper in Arxiv., Comment: 14 pages, 1 figure

The probabilistic investigation on record values and record times of a sequence of random variables defined on the same probability space has received much attention from 1952 to now. A great deal of such theory focused on \textit{iid} or independent real-valued random variables. There exists a few results for real-valued dependent random variables. Some papers deal also with multivariate random variables. But a large theory regarding vectors and dependent data has yet to be done. In preparation of that, the probability laws of records are investigated here, without any assumption on the dependence structure. The results are extended sequences with values in partially ordered spaces whose order is compatible with measurability. The general characterizations are checked in known cases mostly for \textit{iid} sequences. The frame is ready for undertaking a vast study of records theory in high dimensions and for types of dependence., Comment: 48 pages

Asymptotic theories on record values and times, including central limit theorems, make sense only if the sequence of records values (and of record times) is infinite. If not, such theories could not even be an option. In this paper, we give necessary and/or sufficient conditions for the finiteness of the number of records. We prove, for example for \textsl{iid} real valued random variable, that strong upper record values are finite if and only if the upper endpoint is finite and is an atom of the common cumulative distribution function. The only asymptotic study left to us concerns the infinite sequence of hitting times of that upper endpoints, which by the way, is the sequence of weak record times. The asymptotic characterizations are made using negative binomial random variables and the dimensional multinomial random variables. Asymptotic comparison in terms of consistency bounds and confidence intervals on the different sequences of hitting times are provide. The example of a binomial random variable is given, Comment: 15 pages, 1 figure

Using a general strong law of large number proved by Sangar\'e and Lo (2015) and the entropy numbers, we provide functional Glivenko-Cantelli (GC) classes for arbitrary stationary real-valued random variables (rrv's). Next, the general results are particularized for different types of dependence (association, $\phi$-mixing, in particular) and compared with available results in the literature.

This paper proposes a new method for estimating the joint probability mass function of a pair of discrete random variables. This estimator is used to construct joint Shannon R\'enyi-Tsallis entropies, and the mutual information estimates of a pair of discrete random variables. Almost sure consistency and central limit Theorems are established. Our theorical results are validated by simulations., Comment: 20 pages, 4 figures

The like-Lebesgue integral of real-valued measurable functions (abbreviated as \textit{RVM-MI})is the most complete and appropriate integration Theory. Integrals are also defined in abstract spaces since Pettis (1938). In particular, Bochner integrals received much interest with very recent researches. It is very commode to use the \textit{RVM-MI} in constructing Bochner integral in Banach or in locally convex spaces. In this simple not, we prove that the Bochner integral and the \textit{RVM-MI} with respect to a finite measure $m$ are the same on $\mathbb{R}$. Applications of that equality may be useful in weak limits on Banach space., Comment: 11 pages

When applying the classical Stone-Weierstrass common version in Probability Theory for example, and in other fields as well, problems may arise if all points of the compact set are not separated. A solution may consist in going back to the proof and finding alternative versions. In this note, we did it and come back with two flexible versions which are easily used for the needs of classical Probability Theory., Comment: 8

Asymptotic laws of records values have usually been investigated as limits in type. In this paper, we use functional representations of the tail of cumulative distribution functions in the extreme value domain of attraction to directly establish asymptotic laws of records value, not necessarily as limits in type. Results beyond the extreme value value domain are provided. Explicit asymptotic laws concerning very usual laws are listed as well. Some of these laws are expected to be used in fitting distribution, Comment: !7 pages

We present some new nonparametric estimators of entropies and we establish almost sure consistency and central limit Theorems for some of the most important entropies in the discrete case. Our theorical results are validated by simulations., Comment: 18 pages, 4 figures

In this paper, we consider the Zenga index, one of the most recent inequality index. We keep the finite-valued original form and address the asymptotic theory. The asymptotic normality is established through a multinomial representation. The Influence function is also given. Th results are simulated and applied to Senegalese data., Comment: 20 pages

Let the matrix operator $L = D\partial_{xx} + q(x)A_0$, with $D = diag(1, \nu)$, $\nu \neq 1$, $q \in L^{\infty} (0, $\pi$)$, and $A_0$ is a Jordan block of order 1. We analyze the boundary null controllability for system $y_t - Ly = 0$. When $ \nu \notin \mathbb{Q} ^*_+ $ and $q(x) = 1$, $x $\in$ (0, $\pi$)$, there exists a family of root vectors of $(L * , D(L *)) $ forming a Riesz basis, moreover, F. Ammar Khodja, A.Benabdallah, M.Gonzalez-Burgos, L.Teresa, show the existence of a minimal time of control depending on condensation of eigenvalues of $(L^* , D(L^*))$. But there exists $q \in L^{\infty} (0, $\pi$)$ such that the family of eigenfunctions of $(L^* , D(L^*))$ is complete but it is not a Riesz basis. In this framework new phenomena arise : simultaneous condensation of eigenvalues and eigenfunctions. We prove the existence of a minimal time $T_0 \in [0, +\infty]$ depending on the condensation of eigenvalues and associated eigenfunctions of $(L^* , D(L^*))$, such that the corresponding system is null controllable at any time $T > T_0$ and is not if $T < T_0$., Comment: ESAIM: Control, Optimisation and Calculus of Variations, EDP Sciences, In press, 27, pp.S29

In this paper we provide the asymptotic theory of the general of $\phi$-divergences measures, which include the most common divergence measures : Renyi and Tsallis families and the Kullback-Leibler measure. We are interested in divergence measures in the discrete case. One sided and two-sided statistical tests are derived as well as symmetrized estimators. Almost sure rates of convergence and asymptotic normality theorem are obtained in the general case, and next particularized for the Renyi and Tsallis families and for the Kullback-Leibler measure as well. Our theorical results are validated by simulations., Comment: 28 pages, 8 figures. arXiv admin note: substantial text overlap with arXiv:1704.04536

Left and right-continuous functions play an important role in Real analysis, especially in Measure Theory and Integration on the real line and in Stochastic processes indexed by a continuous real time. Semi-continuous functions are also of major interest in the same way. This paper aims at presenting a useful handling of semi-continuous function in parallel with the way left or right continuous functions are treated. For example, a lower or upper continuous function shares the property that it is measurable if it fails to be upper or lower continuous at most at a number of countable points with a left or right-continuous when it fails to be left or right-continuous at most at a number of countable points. As a final result, the comparison between the Riemann and the Lebesgue integrals on compact sets is done in a very comfortable and comprehensive way. The frame used here is open to more further sophistication, Comment: 12

The new Sine Skewed Cardioid (ssc) distribution been just introduced and characterized by Ahsanullah (2018). Here, we study the asymptotic properties of its tails by determining its extreme value domain, the characteristic function, the moments and likelihood estimators of the two parameters, the asymptotic normality of the moments estimators and the random generation of data from the \textit{ssc} distribution. Finally, we proceed to a simulation study to show the performance of the random generation method and the quality of the moments estimation of the parameters.

The univariate extreme value theory deals with the convergence in type of powers of elements of sequences of cumulative distribution functions on the real line when the power index gets infinite. In terms of convergence of random variables, this amounts to the the weak convergence, in the sense of probability measures weak convergence, of the partial maximas of a sequence of independent and identically distributed random variables. In this monograph, this theory is comprehensively studied in the broad frame of weak convergence of random vectors as exposed in Lo et al.(2016). It has two main parts. The first is devoted to its nice mathematical foundation. Most of the materials of this part is taken from the most essential Lo\`eve(1936,177) and Haan (1970), based on the stunning theory of regular, pi or gamma variation. To prepare the statistical applications, a number contributions I made in my PhD and my Doctorate of Sciences are added in the last chapter of the last chapter of that part. Our real concern is to put these materials together with others, among them those of the authors from his PhD dissertations and Science doctorate thesis, in a way to have an almost full coverage of the theory on the real line that may serve as a master course of one semester in our universities. As well, it will help the second part of the monograph. This second part will deal with statistical estimations problems related to extreme values. It addresses various estimation questions and should be considered as the beginning of a survey study to be updated progressively. Research questions are tackled therein. Many results of the author, either unpublished or not sufficiently known, are stated and/or updated therein., Comment: 175 pages

In the footsteps of the book \textit{Measure Theory and Integration By and For the Learner} of our series in Probability Theory and Statistics, we intended to devote a special volume of the very probabilistic aspects of the first cited theory. The book might have assigned the title : From Measure Theory and Integration to Probability Theory. The fundamental aspects of Probability Theory, as described by the keywords and phrases below, are presented, not from experiences as in the book \textit{A Course on Elementary Probability Theory}, but from a pure mathematical view based on Measure Theory. Such an approach places Probability Theory in its natural frame of Functional Analysis and constitutes a firm preparation to the study of Random Analysis and Stochastic processes. At the same time, it offers a solid basis towards Mathematical Statistics Theory. The book will be continuously updated and improved on a yearly basis., Comment: 378 pages

On one hand, a large class of inequality measures, which includes the generalized entropy, the Atkinson, the Gini, etc., for example, has been introduced in Mergane and Lo (2013). On the other hand, the influence function of statistics is an important tool in the asymptotics of a nonparametric statistic. This function has been and is being determined and analysed in various aspects for a large number of statistics. We proceed to a unifying study of the IF of all the members of the so-called Theil-like family and regroup those IF's in one formula. Comparative studies become easier., Comment: 15 pages

In this paper, we consider a coherent theory about the asymptotic representations for a family of inequality indices called Theil-Like Inequality Measures (TLIM), within a Gaussian field. The theory uses the functional empirical process approach. We provide the finite-distribution and uniform asymptotic normality of the elements of the TLIM class in a unified approach rather than in a case by case one. The results are then applied to some UEMOA countries databases., Comment: 19 pages, 4 tables

In this research monograph, we deal with a very general asymptotic representation for statistics named GRI expressed in the functional empirical process, both one-dimensional and multidimensional, and another call residual empirical process. Most of statistics in form of combination of L-statistics are covered by the asymptotic theory dealt here. This treatise is conceived to be a kind of \textbf{spaceship} on which modules are hanged. The spaceship is a functional Gaussian process and each module is the asymptotic representation of one statistic in terms of that Gaussian process. In that way, it is possible to navigate from one module to another, that is, to find the joint distribution of any pair of statistics, to compare them with respect to the areas and the times. In order to be able to do so, we should have a broad conception at the beginning. Within the constructed frame, the asymptotic joint law of any finite number of other statistics is automatically given as well as the joint distribution of its spatial variation or temporal variation, in absolute or relative values. We also deal with the general problem of decomposability of statistics by comparing statistical decomposability, a new view we introduce, versus functional decomposability. A general result only based on the GRI is provided. \noindent This monograph is also the portal of a handbook of GRI that will cover the largest number possible of statistics. In prevision of that, we treat three important examples as show cases. It is expected that this portal and the handbook will attract the attention of researchers working in the asymptotic area and will furnish useful tools to scientists who are interested in application of asymptotic tests, completed by computer packages.

In this note we provide a quick proof of the Sklar's Theorem on the existence of copulas by using the generalized inverse functions as in the one dimensional case, but a little more sophisticated., Comment: 5

We describe the main features of the households databases we can find in most of our National Statistics Institute. We provide algorithms aimed at extracting a diversity of variables on which different statistical procedures may be applied. Here, we particularly focus on the scaled income, as a beginning. Associated codes (MS Visual Basic and R codes) have been successfully tested and delivered in the text and in a separate file, Comment: 42 pages, 3 figures

A number of methods have been introduced in order to measure the inequality in various situations such as income and expenditure. In order to curry out statistical inference, one often needs to estimate the available measures of inequality. Many estimators are available in the literature, the most used ones being the non parametric estimators. kpanzou(2011) has developed semi-parametric estimators for measures of inequality and showed that these are very appropriate especially for heavy tailed distributions. In this paper we apply such semi-parametric methods to a practical data set and show how they compare to the non parametric estimators. A guidance is also given on the choice of parametric distributions to fit in the tails of the data, Comment: 10

Uniform convergence rates are provided for asymptotic representations of sample extremes. These bounds which are universal in the sense that they do not depend on the extreme value index are meant to be extended to arbitrary samples extremes in coming papers., Comment: 10

This paper introduces a general continuous form of poverty index that encompasses most of the existing formulas in the literature. We then propose a consistent estimator for this index in case the poverty line is a functional of the distribution. We also establish a uniform functional Central Limit Theorem for the proposed estimator over a suitable product class of real-valued functions. As a consequence, testing procedures based either on single or simultaneously several poverty indices can be developed. A simulation study showing the asymptotic normality of the estimator is given as well as an application to real data for estimating the effect of relative poverty lines on the variance of the poverty estimates.

Measure Theory and Integration is exposed with the clear aim to help beginning learners to perfectly master its essence. In opposition of a delivery of the contents in an academic and vertical course, the knowledge is broken into exercises which are left to the learners for solutions. Hints are present at any corner to help readers to achieve the solutions. In that way, the knowledge is constructed by the readers by summarizing the results of one or a group of exercises. Each chapter is organized into Summary documents which contain the knowledge, Discovery documents which give the learner the opportunity to extract the knowledge himself through exercises and into Solution Documents which offer detailed answers for the exercises. Exceptionally, a few number of results (A key lemma related the justification of definition of the integral of a non-negative function, the Caratheodory's theorem and the Lebesgue-Stieljes measure on $\mathbb{R}^d$) are presented in appendix documents and given for reading in small groups. The full theory is presented in the described way. We highly expect that any student who goes through the materials, alone or in a small group or under the supervision of an assistant will gain a very solid knowledge in the subject and by the way ensure a sound foundation for studying disciplines such as Probability Theory, Statistics, Functional Analysis, etc. The materials have been successfully used as such in normal real analysis classes several times., Comment: 715 pages

In this paper we provide the asymptotic theory of the general of $\phi$-divergences measures, which includes the most common divergence measures : Renyi and Tsallis families and the Kullback-Leibler measure. Instead of using the Parzen nonparametric estimators of the probability density functions whose discrepancy is estimated, we use the wavelets approach and the geometry of Besov spaces. One-sided and two-sided statistical tests are derived as well as symmetrized estimators. Almost sure rates of convergence and asymptotic normality theorem are obtained in the general case, and next particularized for the Renyi and Tsallis families and for the Kullback-Leibler measure as well. The applicability of the results to usual distribution functions is addressed., Comment: 33 pages