Coefficients de corrélation d'ordre supérieur
In this paper we give the expression of the multiple correlation coefficient in a linear model according to the coefficients of correlation. This expression makes it possible to analyze from a numerical point of view the instability of estimates in the case of collinear explanatory variables in the linear model or in the autoregressive model. This numerical approach, that we show on two examples, thus supplements the usual approach of the quasi colinearity, founded on the statistical properties...
Prototype Selection (PS) techniques have traditionally been applied prior to Nearest Neighbour (NN) classification rules both to improve its accuracy (editing) and to alleviate its computational burden (condensing). Methods based on selecting/discarding prototypes and methods based on adapting prototypes have been separately introduced to deal with this problem. Different approaches to this problem are considered in this paper and their main advantages and drawbacks are pointed out along with some...
Meta-analysis is a standard statistical method used to combine the conclusions of individual studies that are related and the results of single study alone can not answered to deal with issues. The data are summarized by one or more outcome measure estimates along with their standard errors. The multivariate model and the variations between studies are not considered in most articles. Here we discuss multivariate effects models: a multivariate fixed effects model and a multivariate random effects...
Comparing q-ary relations on a set of elementary objects is one of the most fundamental problems of classification and combinatorial data analysis. In this paper the specific comparison task that involves classification tree structures (binary or not) is considered in this context. Two mathematical representations are proposed. One is defined in terms of a weighted binary relation; the second uses a 4-ary relation. The most classical approaches to tree comparison are discussed in the context...
Comparing q-ary relations on a set of elementary objects is one of the most fundamental problems of classification and combinatorial data analysis. In this paper the specific comparison task that involves classification tree structures (binary or not) is considered in this context. Two mathematical representations are proposed. One is defined in terms of a weighted binary relation; the second uses a 4-ary relation. The most classical approaches to tree comparison are discussed in the context...
We will discuss orthogonal models and error orthogonal models and their algebraic structure, using as background, commutative Jordan algebras. The role of perfect families of symmetric matrices will be emphasized, since they will play an important part in the construction of the estimators for the relevant parameters. Perfect families of symmetric matrices form a basis for the commutative Jordan algebra they generate. When normality is assumed, these perfect families of symmetric matrices will ensure...
The class of componentwise concave copulas is considered, with particular emphasis on its closure under some constructions of copulas (e.g., ordinal sum) and its relations with other classes of copulas characterized by some notions of concavity and/or convexity. Then, a sharp upper bound is given for the -measure of non-exchangeability for copulas belonging to this class.
Data depth is an important concept of nonparametric approach to multivariate data analysis. The main aim of the paper is to review possible applications of the data depth, including outlier detection, robust and affine-equivariant estimates of location, rank tests for multivariate scale difference, control charts for multivariate processes, and depth-based classifiers solving discrimination problem.
Cette étude offre une large vision de synthèse prospective ; mais aussi, des résultats techniques précis sur une famille très générale que nous avons élaborée de coefficients d'association entre variables descriptives relationnelles à partir de leur observation empirique sur un ensemble O d'objets élémentaires. Un même coefficient est obtenu à partir d'une forme de normalisation statistique par rapport à une hypothèse d'absence de liaison, d'un indice brut d'association. Ce dernier suppose une représentation...
Cette étude offre une large vision de synthèse prospective : mais aussi, des résultats techniques précis sur une famille très générale que nous avons élaborée de coefficients d'association entre variables descriptives relationnelles à partir de leur observation empirique sur un ensemble O d'objets élémentaires. Un même coefficient est obtenu à partir d'une forme de normalisation statistique par rapport à une hypothèse d'absence de liaison, d'un indice brut d'association. Ce dernier suppose une représentation...
We consider here a multivariate sample Xj = (X1.j > ... > Xi.j), 1 ≤ j ≤ n, where the Xj, 1 ≤ j ≤ n, are independent i-dimensional extremal vectors with suitable unknown location and scale parameters λ and δ respectively. Being interested in linear estimation of these parameters, we consider the multivariate sample Zj, 1 ≤ j ≤ n, of the order statistic of largest values and their concomitants, and the best linear unbiased estimators of λ and δ based on such multivariate sample. Computational...