RT info:eu-repo/semantics/other
T1 An R script to compute the weighted right/left tail index and a new skewness functional for univariate probability distributions
A1 Arriaza Gómez, Antonio Jesús
A1 Sordo, Miguel Ángel
A1 Di Crescenzo, Antonio
A1 Suárez-Llorens, Alfonso
A2 Estadística e Investigación Operativa
K1 shape measures
K1 skewness measures
K1 convex transform order
K1 kurtosis and skewness
K1 heavy tail
AB An R script to compute the weighted left and right tail indices as well as the skewness functional provided in the paper entitle: "Skewness and kurtosis measures based on the convex transform order". Paper submitted to Metrika Journal (November 2017). We provide along with this contribution a help document (ReadMe.pdf).
YR 2017
FD 2017-11-16
LK http://hdl.handle.net/10498/19833
UL http://hdl.handle.net/10498/19833
LA eng
NO The weighted left and right tail indices are new shape measures for univariate probability distributions, provided in "Skewness and kurtosis measures based on the convex transform order" (submitted to Metrika Journal), and consistent with the convex transform order defined by van Zwet in the book entitle "Convex transformations of random variables" (1964). These new indices allow us to characterize location-scale families. Furthermore, they can be use to compare two probability distributions in terms of tail weight. On the other hand, both measures are used to define a new skewness functional which characterizes symmetry distributions. The tailMeasures.R file is an script written in R, which let us to study these measures for a random variable X. It works computing numerically the weighted left and right tail measures (denoted by Lx(u) and Rx(u), respectively) from a density and a quantile functions provided by the user. From both functions, the program calculates the skewness functional SKx(u) which compares the weight of both tails (left and right) of the probability distribution associated to X. Finally, this script provides the plots of Lx(u), Rx(u) and SKx(u). Additionally, the plots are saved in the local folder as pdf files.
DS Repositorio Institucional de la Universidad de Cádiz
RD 25-sep-2020