Publication: A quantile approach to the box-cox transformation in random samples
Loading...
Identifiers
Publication date
1991-03
Defense date
Authors
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Impact
Export
Abstract
This paper presents an alternative approach to the likelihood methods for estimating the parameter A in the Box-Cox family of transformations when the data arise from a random sample. The method is based on a representation of the quantile function of the variable under consideration. Theoretical properties of the method, its practical applications and comparison with the likelihood approach are studied.
Description
Keywords
Asymptotic relative efficiency (ARE), Box-Cox transformation, Influential observations, Quantile function, Kernel density estimation