- Tytuł:
- Measuring and Testing Mutual Dependence of Multivariate Functional Data
- Autorzy:
-
Krzyśko, Mirosław
Smaga, Łukasz - Powiązania:
- https://bibliotekanauki.pl/articles/1058987.pdf
- Data publikacji:
- 2020-09-04
- Wydawca:
- Główny Urząd Statystyczny
- Tematy:
-
characteristic function
dependence measure
distance covariance
multivariate functional data
permutation method
test of independence - Opis:
- This paper considers new measures of mutual dependence between multiple multivariate random processes representing multidimensional functional data. In the case of two processes, the extension of functional distance correlation is used by selecting appropriate weight function in the weighted distance between characteristic functions of joint and marginal distributions. For multiple random processes, two measures are sums of squared measures for pairwise dependence. The dependence measures are zero if and only if the random processes are mutually independent. This property is used to construct permutation tests for mutual independence of random processes. The finite sample properties of these tests are investigated in simulation studies. The use of the tests and the results of simulation studies are illustrated with an example based on real data.
- Źródło:
-
Statistics in Transition new series; 2020, 21, 3; 21-37
1234-7655 - Pojawia się w:
- Statistics in Transition new series
- Dostawca treści:
- Biblioteka Nauki