CHANGES

pareto Changelog
================

Here you can see the full list of changes between each pareto release.

Version 0.3
-----------

Released on March 29th, 2016

- Added MLE for 'Gamma' distribution.
- Added log-probability and log-density to all distributions.
- Added a method for sampling a single value from a distribution, see
  'BaseDistribution.random'.
- Changed the observation label in 'DiscreteDistribution' from ~n to ~k.
- Bumped minimum required ocaml-gsl version to 1.14.0 and dropped custom
  'Combi' stubs.

Version 0.2
-----------

Released on July 13th, 2013

- Added summary statistics, see 'Sample.Summary' module. Thanks to Nicholas
  Lucaroni!
- Added MLE for distributions which have a closed-form MLE. Thanks to Nicholas
  Lucaroni, once again! For distributions, which don't have a closed-form
  solution, we currently provide method of moments estimates. Exceptions are:
  'Cauchy' and 'Hypergeomteric' distributions.
- Added Pearson product-moment correlation and autocorrelation and Spearman
  rank correlation.
- Added Kolmogorov-Smirnov test for goodness of fit and two-sample test.
- Added function, which adjusts P-values for multiple comparisons. Currently,
  two procedures are supported: Benjamini-Hochberg, which controls for FDR and
  Holm-Bonferroni, which controls FWER.
- Abstracted RV type into '*Distribution.elt'.
- Added GSL wrappers for computing sample skewness and kurtosis.
- Added more distribution features: skewness and kurtosis, see
  'Distributions.Features' signature.
- Added 'LogNormal', 'Bernoulli', 'Logistic' and 'Categorical' distributions.
- Added a basic test suite against R and SciPy, thanks to Francois Berenger
  for the reminder.

Version 0.1
-----------

Initial release, released on June 11th, 2013