On application of the asymptotic tests for estimating the number of mixture distribution components

On application of the asymptotic tests for estimating the number of mixture distribution components

The paper demonstrates the efficiency of asymptotically most powerful test of statistical hypotheses about the number of mixture components in the adding and splitting component models. Test data are the samples from different finite normal mixtures. The results are compared for various significance...

Full description

Bibliographic Details
Main Author: Andrey Konstantinovich Gorshenin
Format: Article
Language:Russian
Published: Institute of Computer Science 2012-03-01
Series:Компьютерные исследования и моделирование
Subjects:
finite mixture of normal distributions
asymptotically most powerful test
Online Access:http://crm.ics.org.ru/uploads/crmissues/crm_2012_1/12103.pdf
id doaj-255c6ce3a3854e33a420f8d4112920a4
record_format Article
spelling doaj-255c6ce3a3854e33a420f8d4112920a42020-11-25T02:17:11ZrusInstitute of Computer ScienceКомпьютерные исследования и моделирование2076-76332077-68532012-03-0141455310.20537/2076-7633-2012-4-1-45-531854On application of the asymptotic tests for estimating the number of mixture distribution componentsAndrey Konstantinovich GorsheninThe paper demonstrates the efficiency of asymptotically most powerful test of statistical hypotheses about the number of mixture components in the adding and splitting component models. Test data are the samples from different finite normal mixtures. The results are compared for various significance levels and weights.http://crm.ics.org.ru/uploads/crmissues/crm_2012_1/12103.pdffinite mixture of normal distributionsasymptotically most powerful test
collection DOAJ
language Russian
format Article
sources DOAJ
author Andrey Konstantinovich Gorshenin
spellingShingle Andrey Konstantinovich Gorshenin
On application of the asymptotic tests for estimating the number of mixture distribution components
Компьютерные исследования и моделирование
finite mixture of normal distributions
asymptotically most powerful test
author_facet Andrey Konstantinovich Gorshenin
author_sort Andrey Konstantinovich Gorshenin
title On application of the asymptotic tests for estimating the number of mixture distribution components
title_short On application of the asymptotic tests for estimating the number of mixture distribution components
title_full On application of the asymptotic tests for estimating the number of mixture distribution components
title_fullStr On application of the asymptotic tests for estimating the number of mixture distribution components
title_full_unstemmed On application of the asymptotic tests for estimating the number of mixture distribution components
title_sort on application of the asymptotic tests for estimating the number of mixture distribution components
publisher Institute of Computer Science
series Компьютерные исследования и моделирование
issn 2076-7633
2077-6853
publishDate 2012-03-01
description The paper demonstrates the efficiency of asymptotically most powerful test of statistical hypotheses about the number of mixture components in the adding and splitting component models. Test data are the samples from different finite normal mixtures. The results are compared for various significance levels and weights.
topic finite mixture of normal distributions
asymptotically most powerful test
url http://crm.ics.org.ru/uploads/crmissues/crm_2012_1/12103.pdf
work_keys_str_mv AT andreykonstantinovichgorshenin onapplicationoftheasymptotictestsforestimatingthenumberofmixturedistributioncomponents
_version_ 1724887700633288704

Similar Items