Temat: measures of dependence

Skocz do pozycji: 1.

Tytuł:: Miltivariate measures of dependence based on copulas
Autorzy:: Heilpern, Stanisław
Tematy:: multivariate measures of dependence
copulas
tail dependences
estimation; Pokaż więcej
Wydawca:: Wydawnictwo Uniwersytetu Ekonomicznego we Wrocławiu
Powiązania:: https://bibliotekanauki.pl/articles/584976.pdf Link otwiera się w nowym oknie
Opis:: The paper is devoted to the multivariate measures of dependence. In contrast to the classical approach, where the pairs of variables are studied, we investigate the dependence of more than two variables. We mainly consider the measures based on copulas. These are the multivariable generalizations of the known coefficients of such correlation as Spearman’s rho, Kendall’s tau, Blomquist’s beta and Gini’s gamma. We present the definitions, the constructions and the basic properties of such multivariate measures of dependence. The case of large number of dimension, greater than two, presents more complications. We have several different versions of such generalization in this case and the lower bound of the values of such measures of dependence are close to zero. We also study the multivariate tail dependences. The last part of the paper is devoted to the estimation of multivariable versions of Spearman’s rho coefficient.
Dostawca treści:: Biblioteka Nauki

Artykuł

Skocz do pozycji: 2.

Tytuł:: Raport Badawczy = Research Report ; RB/13/2012
On computational problems in the analysis of statistical dependence for imprecise data
Autorzy:: Hryniewicz, Olgierd (1948– )
Opara, Karol
Wydawca:: Instytut Badań Systemowych. Polska Akademia Nauk
Systems Research Institute. Polish Academy of Sciences
Powiązania:: Raport Badawczy = Research Report
Opis:: Bibliografia s. 38-40
Bibliography p. 38-40
This paper provides a comprehensive analysis of computational problems concerning calculation of Kendall’s τ dependence measure for interval data. Exact algorithms solving this task have unacceptable computational complexity for larger samples, therefore the focus is on computational problems arising in approximate algorithms. In particular, a set of heuristic solutions was proposed to find minimal and maximal value of Kendall’s τ for interval data. Extensive simulation experiments show that some of the heuristics yield very good starting points for optimization procedures based on random generation of linear extensions or evolutionary and direct-search algorithms.
40 pages ; 21 cm
40 stron ; 21 cm
Dostawca treści:: RCIN - Repozytorium Cyfrowe Instytutów Naukowych

Książka

Skocz do pozycji: 3.

Tytuł:: Prediction of reliability – pitfalls of using Pearson’s correlation
Raport Badawczy = Research Report ; RB/19/2012
Autorzy:: Hryniewicz, Olgierd (1948– )
Karpiński, Janusz
Wydawca:: Instytut Badań Systemowych. Polska Akademia Nauk
Systems Research Institute. Polish Academy of Sciences
Powiązania:: Raport Badawczy = Research Report
Opis:: [29] stron ; 21 cm
Pearson’s coefficient of linear correlation r is the most popular among practitioners measure of dependence. This paper presents, using comprehensive computer simulations, that its application is very limited when we search for informative variables that can be used for the prediction of reliability. Kendall’s τ coefficient has been shown to be much better for this purpose.
Bibliography p. [28-29]
Bibliografia s. [28-29]
[29] pages ; 21 cm
Dostawca treści:: RCIN - Repozytorium Cyfrowe Instytutów Naukowych

Książka

Skocz do pozycji: 4.

Tytuł:: Estimators of g-monotone dependence functions
Autorzy:: Krajka, Andrzej
Tematy:: sample
monotone dependence functions
measures of dependence
strong law of large numbers
quantile
independent
positively (negatively) quadrant dependent random variables; Pokaż więcej
Wydawca:: Polska Akademia Nauk. Instytut Matematyczny PAN
Powiązania:: https://bibliotekanauki.pl/articles/1338960.pdf Link otwiera się w nowym oknie
Opis:: The notion of g-monotone dependence function introduced in [4] generalizes the notions of the monotone dependence function and the quantile monotone dependence function defined in [2], [3] and [6]. In this paper we study the asymptotic behaviour of sample g-monotone dependence functions and their strong properties.
Dostawca treści:: Biblioteka Nauki

Artykuł

Skocz do pozycji: 5.

Książka

Informacja