EXTREME: Stata module to fit models used in univariate extreme value theory
extreme estimates the most-used models in univariate extreme value theory, using Maximum Likelihood: the generalized Pareto distribution (GPD), which is appropriate for modeling exceedances of a threshold; the generalized extreme value distribution (GEV), which is appropriate for modeling block maxima; and the extension of the GEV for multiple order statistics for blocks. extreme also provides a variety of diagnostic and profile plots. extreme's major novelty is the ability to compute the Cox-Snell small-sample bias correction for all models fit. The correction for the GPD is derived in Giles, Feng, and Godwin (2015). The correction for the GEV, including for multiple order statistics, is new. The correction is also extended to non-stationary models. Maximum Likelihood is always biased in finite samples, and the bias can be significant in the small samples often used in extreme value analysis.
|Requires:||Stata version 11|
|Date of creation:||18 Jan 2015|
|Date of revision:||18 Dec 2017|
|Note:||This module should be installed from within Stata by typing "ssc install extreme". Windows users should not attempt to download these files with a web browser.|
|Contact details of provider:|| Postal: Boston College, 140 Commonwealth Avenue, Chestnut Hill MA 02467 USA|
Web page: http://fmwww.bc.edu/EC/
More information through EDIRC
|Order Information:||Web: http://repec.org/docs/ssc.php|
When requesting a correction, please mention this item's handle: RePEc:boc:bocode:s457953. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Christopher F Baum)
If references are entirely missing, you can add them using this form.