Search

Your search keyword '"Kullback–Leibler divergence"' showing total 3 results
Start Over Descriptor "Kullback–Leibler divergence" Language turkish
3 results on '"Kullback–Leibler divergence"'

1. Entropi, İnformasyon ve Entropi Korelasyon Katsayısı.

Author

KAYA, Mehmet Onur, GÜRAL, Yunus, and GÜRCAN, Mehmet

Subjects
*STATISTICAL correlation, *ENTROPY (Information theory), *DATA structures, *PROBABILITY theory
Abstract

Statistical analyzes are methods that help the researcher to interpret correctly and increase the amount of information obtained by the researcher. Researchers may be undecided in interpreting the information they obtain from observations. It is important to recommend methods that can facilitate researchers in determining this uncertainty. In this way, the researcher can positively change the design of the study. Entropy and information are important concepts in terms of showing researchers the amount of information they obtain after observation or a measure of the irregularity in the data. Objective: The aim of the study is to interpret the entropy correlation coefficient by evaluating the concepts of entropy and information in terms of probability. Material and Methods: In most applications, probability is used only in the interpretation of relative frequencies in the observed data. Relative frequencies are not used for entropy calculation. The main method to be used in the study is to determine the possible relationships between the variables identified with the help of observed events and to obtain their correlations. Results: The first of the two important findings obtained in the study is to determine the entropy of the regression error term, and the second is to find a correlation of the variables defining the common entropies among the four variables with weak correlations. Conclusion: In the multivariate data structure, the linear form of some variables may be related to the linear form of some other variables. In this case, the correlation of the two linear forms will be high. It is possible to detect this by creating linear forms. In such a case, the entropy correlation coefficient can be easily used. [ABSTRACT FROM AUTHOR]

Published
2020
Full Text
View/download PDF

3. Goodness of fit tests based on Kullback-Leibler divergence under censoring

Author

Gencer, Gülcan, Kınacı, İsmail, and Enstitüler, Fen Bilimleri Enstitüsü, İstatistik Ana Bilim Dalı

Subjects
Censoring, Goodness of fit test, Kullback-Leibler uyuşmazlığı, Bandwidth, Bant genişliği, Kullback-Leibler divergence, Test statistics, Sansürleme, Uyum iyiliği testi, Test istatistikleri
Abstract

Bilinen bir teorik dağılım ile örnek değer içeren dağılım arasındaki uyumun ölçülmesi uyum iyiliği testi olarak adlandırılır. Bu tez çalışmasında, sansürlü ve tam örneklem durumunda Kullback-Leibler uyuşmazlığına dayalı çeşitli sürekli dağılım aileleri için uyum iyiliği testleri en iyi ve en kötü bant genişliği baz alınarak kritik değerler ve güçler elde edilmiştir. İlk bölümde, Uyum iyiliği testleri tanımlanmış ve Kullback-Leibler uyuşmazlığı, ayırma problemi hakıında literatür taraması verilmiştir. İkinci bölümde, tezde kullanılan temel kavramlar açıklanmıştır. Üçüncü bölümde bazı dağılımlar ile ilgili olarak tam ve sansürlü örneklem durumunda Kullback-Leibler uyuşmazlığına dayalı uyum iyiliği testleri yapılmıştır. Bu bölümlerde ele alınan dağılımlar farklı parametre durumları için genişletilmiştir. Dördüncü bölümde, üçüncü bölümde elde edilen sonuçlara ilişkin genel yorumlar yapılarak, tez için öneriler sunulmuştur., Measuring the fit between a known theoretical distribution and the distribution containing sample value is called the goodness of fit test. In this thesis study, the critical values and powers were obtained based on the best and worst bandwidth tests for various continuous distribution families based on Kullback-Leibler divergence in censored and complete sample situations. In the first part, the goodness of fit tests are defined and a literature review is given about Kullback-Leibler divergence and discriminant problem. In the second chapter, the basic concepts used in the thesis are explained. In the third part, goodness of fit tests based on Kullback-Leibler divergence were conducted for some distributions in case of complete and censored samples. The distributions discussed in these sections are expanded for different parameter cases. In the fourth chapter, general comments about the results obtained in the third chapter are made and suggestions for the thesis are presented.

Published
2020

Catalog

Books, media, physical & digital resources
See catalog results