Result Details
Asymptotic Tests for Multiply Left-Censored Samples from Weibull Distribution
Michálek Jaroslav, doc. RNDr., CSc., FME (FME), UTKO (FEEC), IM (FME)
Left-censored data with one or more detection limits occur frequently in many application areas. This paper suggests the computational procedure for calculation of maximum likelihood estimates of the parameters and estimates of their variances for type I multiply left-censored Weibull samples. Estimates of the variances of estimated parameters are based on the analytically determined expected Fisher information matrix. Moreover, using the asymptotic properties of maximum likelihood estimates and tests with nuisance parameters (Lagrange multiplier test, likelihood ratio test, Wald test), methods for comparison of two independent type I multiply left-censored Weibull samples are proposed. The power functions of particular tests are compared by simulations. The methods derived in this paper can be used in real environmental or chemical data analysis.
Fisher information matrix, maximum likelihood, musk compounds, power of test
@inproceedings{BUT108106,
author="Michal {Fusek} and Jaroslav {Michálek}",
title="Asymptotic Tests for Multiply Left-Censored Samples from Weibull Distribution",
booktitle="MENDEL 2014, 20th International Conference on Soft Computing",
year="2014",
journal="Mendel Journal series",
pages="317--322",
publisher="Brno University of Technology, Faculty of Mechanical Engineering, Institute of Automation and Computer Science",
address="Brno, Czech Republic",
isbn="978-80-214-4984-8",
issn="1803-3814"
}