Atbilstības testi atkarīgiem novērojumiem
Loading...
Date
Authors
Advisor
Journal Title
Journal ISSN
Volume Title
Publisher
Latvijas Universitāte
Language
N/A
Abstract
Praksē plaši tiek pielietoti testi par sadalījumu neatkarīgiem un vienādi sadalītiem novērojumiem, piemēram, klasiskais Hī-kvadrāta vai Kolmogorova-Smirnova tests. Tomēr ļoti bieži dati ir atkarīgi. Jauktiem procesiem, kas vispārīgi apraksta dažādas atkarības struktūras 2000.gadā tika ieviests Bikela-Rozenblata tests. Darbā tiek veikta šī testa teorētiska izpēte un empīriskās jaudas simulāciju analīze. Rezultāti tiek salīdzināti ar Neimaņa testu, kas jauktiem procesiem savukārt tika ieviests samērā nesen.
In practice there are commonly applied tests about distribution for independent and identically distributed observations, for example, the classical Chi-square or Kolmogorov-Smirnov tests. However, very often data are dependent. For mixing processes, which in general describe dependence structure Bickel-Rosenblatt test has been established in 2000. In this work theoretical investigation and analysis of the empirical power for this test are considered. Results are compared with the Neyman test, which has been introduced relatively recently for mixing processes.
In practice there are commonly applied tests about distribution for independent and identically distributed observations, for example, the classical Chi-square or Kolmogorov-Smirnov tests. However, very often data are dependent. For mixing processes, which in general describe dependence structure Bickel-Rosenblatt test has been established in 2000. In this work theoretical investigation and analysis of the empirical power for this test are considered. Results are compared with the Neyman test, which has been introduced relatively recently for mixing processes.