Between the sampling distribution of CVc and CVr for exponential samples from simulated and real data.
First part of this paper reviews the results, obtained via simulation, of the properties of the sampling distribution of the conventional coefficient of variation, CVc and that of a robust coefficient of variation based on the mean absolute deviation (MAD), CVr for samples from exponential dist...
| Main Authors: | , , |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | English English |
| Online Access: | http://psasir.upm.edu.my/id/eprint/27731/ http://psasir.upm.edu.my/id/eprint/27731/1/ID%2027731.pdf |
| Summary: | First part of this paper reviews the results, obtained
via simulation, of the properties of the sampling distribution of
the conventional coefficient of variation, CVc and that of a
robust coefficient of variation based on the mean absolute
deviation (MAD), CVr for samples from exponential
distribution. The properties of emphasis are symmetricity,
peakedness and spread. Second part of the paper presents
exemplification on the preservation of the observed properties
in exponentially distributed engineering relayed real data sets.
Also presented are observations made on the effect of
stipulated presence of outliers in the real data sets used. |
|---|