DistributionFitR (version 0.1)

DistributionFitR-package: Fitting Multiple Parametric Distributions

Description

Parametric densities (or count densities) will be fitted to user-given univariate data via maximum likelihood. The user more or less only enters the data. The program automatically searches for parametric distributions and parameters thereof that best describe the data. It then returns the ten best parameter families including the fitted parameters.

DistributionFitR comes in with a standard search list of 408 parametric distribution families as given by R-packages on CRAN.

The package contains the following functions:

• globalfit Given some univariate data, determines the best fitting parametric distributions from R-packages

• install.packages_DistributionFitR Installs R-packages that contribute to DistributionFitR's standard search list

• getFamilies Browsing for distribution families

• getFamily Find all distributions within a package

• getParams For a single distribution family, parameters are determined; upper/lower limits and other chraracteristics are extracted

References

• Zucchini, W. (2000). An introduction to model selection. Journal of Mathematical Psychology, 44(1), 41-61.

• Shao, J. (2003) Mathematical Statistics New York: Springer, ISBN 978-0-387-21718-5.

See Also

globalfit, install.packages_DistributionFitR, getFamilies, getFamily, getParams

Examples

# NOT RUN {
# example for globalfit
data <- rnorm(n = 100, mean = 70, sd = 4)
r <- globalfit(data, cores = if(interactive()) NULL else 2)
summary(r)

# example for getFamily
str(getFamily("stats"))

# example for getParams
getParams("beta", package = "stats")

# }