ROOT ESTIMATION OF DENSITY FUNCTION USING INCOMPLETE DATA
https://doi.org/10.21683/1729-2646-2013-0-4-44-63
Abstract
The paper offers two modifications of the nonparametric root estimation of a density function in case of incomplete data in the form of grouped frequencies of failures. The first (integrated) method is connected with respective alteration of a likelihood function. The second (resampling) method of the restoration of failures is based on the iterative restoration of failure time points. The accuracy of the offered estimation methods has been investigated.
About the Authors
E. A. Maleev
Obninsk Institute of Atomic Energy, National Research Nuclear University MIFI
Russian Federation
Post graduate student of chair of automated control systems
Russian Federation
Post graduate student of chair of automated control systems
V. A. Chepurko
Obninsk Institute of Atomic Energy, National Research Nuclear University MIFI
Russian Federation
PhD, associate professor of chair of automated control systems
Russian Federation
PhD, associate professor of chair of automated control systems
References
1. AntonovA.V., Chepurko V.A. Definition of nonparametric density function on the basis of censored data. Reliability. – М.: The Publishing house “Technology”, 2005, №2. – p.3.
2. Bogdanov U.I. The primary task of statistical analysis of data: the root approach. – M: МIET, 2002. – 96 p.
3. KryanevA.V., Lukin G.V. Mathematical methods of uncertain data processing. – М.: PHYSMATHLIT, 2003. – 216 p.
4. Ershov A.N., Chepurko V.A. Iterative estimation of parameters of the distribution law of a random variable at availability of censored data. Diagnostics and forecasting of complex systems’ state: the collection of proceedings № 18 каф. Chair of Automatic Control Systems – Obninsk: Institute for Nuclear Power Engineering, 2009. p. 14-22.
Review
For citations:
Maleev E.A., Chepurko V.A. ROOT ESTIMATION OF DENSITY FUNCTION USING INCOMPLETE DATA. Dependability. 2013;(4):44-63. https://doi.org/10.21683/1729-2646-2013-0-4-44-63
Views:
514
ISSN 1729-2646 (Print)
ISSN 2500-3909 (Online)
ISSN 2500-3909 (Online)
Email this article 