Changes in version 2.0.1 (2026-02-15) o Version packaged with R-4.5.2 and sent to CRAN (R-4.5.2) o fatreturns.R: change timeSeries::time(z) <- timeSeries::time(x) to timeSeries::time(z) <- stat::time(x) as per Georgi Boshnakov's recommandation to cope with the next version of timeSeries Changes in version 2.0.0 (2025-04-21) o Version packaged with R-4.5.0 and sent to CRAN (R-4.5.0) o First version on CRAN with the scaling parameter (x-m)*pi/sqrt(3)/g to get sd=1 at k=Inf o BREAK with previous results for the g parameter (other parameters stay unchanged). Changes in version 1.9.10 o version 1.9.9 packaged with R-4.5.0 o Depends changed from R >= 3.1.0 to R >= 4.1.0 o FatTailsR 1.9.11: 2025-04-20 change the DESCRIPTION Changes in version 1.9.9 o FatTailsR 1.9.9: rewrite all equations in proper Latex format with deqn{}{} and eqn{}{} o FatTailsR 1.9.8: rewrite kiener1234.R examples using outer() o FatTailsR 1.9.7: rewrite pkiener2347: 20x faster and much more accurate (= BIG IMPROVEMENT !!) has a similar impact on dkiener2347: 20x faster and more accurate update the examples in kiener1234 with new values for sd, xylim, logisst rewrite the examples in kashp o FatTailsR 1.9.6: update the the doc (kashp,k12347,logisst) with the new conversion dimdim is now an internal function with a hidden doc o FatTailsR 1.9.5: replace *logishp by *logisst (standardized logistic distribution). start rewriting the doc o FatTailsR 1.9.4: moments. Conversion from (x-m)/2/g to (x-m)*pi/sqrt(3)/g seems OK everywhere o FatTailsR 1.9.3: moments o FatTailsR 1.9.3: dpqr++logishp, ltm,rtmlogis, hkiener12347, hfitkX:logisk o FatTailsR 1.9.2: rtm, ltm o FatTailsR 1.9.1: estimkiener11, .hparamkienerX5 Changes in version 1.9.0 o Start the transition to v2 with the scaling parameter (x-m)*pi/sqrt(3)/g rather than (x-m)/2/g in order to get a variance of 1 when kappa = Inf o Remove the trigonometric functions (see them in FatTailsR23) Changes in version 1.8-6 o version packaged for R-4.4.0 o FINAL INTERNAL VERSION WITH (x-m)/2/g Changes in version 1.8-5 (2024-03-03) o Version packaged with R-4.3.2 and sent to CRAN (R-4.3.3) o !!! FINAL PUBLIC VERSION ON CRAN WITH (x-m)/2/g !!! Changes in version 1.8-4 o Prepared for R-4.3.2 o rewrite 2x {p} as \code{p} in f_kiener12347 to pass new checks o replace \item{}{} by \item in m_laplaceroll to pass new checks o rename FatTailsR-package.R and arrange @include in b_data.R o split b_data.R in a_data.R and b_fatreturns.R. Arrange the @include o DESCRIPTION: arrange collate Changes in version 1.8-3 o Prepared for R-4.2.2 Changes in version 1.8-2 o Prepared with R-4.2.0 Changes in version 1.8-0 (2021-03-12) o Prepared for R-4.1.0 o 36 http links changed to https thanks to urlchecker package o Roxygen 7.1.1 Changes in version 1.7-6 o paramkienerX5 and X7 rewritten : new algorithm -- nonlinear interpolation rather than second order approximation Changes in version 1.7-5 (2017-05-16) o Codename: Sister's birthday o Version for MNHN Semin-R conference (2017-06-16) -- https://rug.mnhn.fr/semin-r/PDF/JourneeR-2017-Programme.pdf o Nothing new in this subversion o MAIN NOVELTIES vs 1.7-0: Changes in version 1.7-3 o mData o extractData rewritten: "zoo" produces an index in Date format Changes in version 1.7-2 o Package generated from Windows and Linux Changes in version 1.7-0 (2016-05-15) o Codename: Dimanche de Pentecote o MAIN NOVELTIES vs 1.6-0: -- (dpqr)kiener7() vectorized for vector and matrix -- checkcoefk(), replaceNA() -- fatreturns(), logret() and elevate() o See details below Changes in version 1.6-6 o Version for AB and documentation synchronised with FatTailsRplot. o checkquantiles(..., STOP = FALSE) in estimkiener5/7/11(). Changes in version 1.6-4 o extractData(pr="r") uses fatreturns rather than 100*diff(log()). Much better... Changes in version 1.6-3 o New page *kiener7.R and new functions *kiener7() with (x,q,p) integer or vector and (coefk) vector or matrix. o New parameter signedES in eskiener12347.R(..., signedES = FALSE) FALSE = (abs(ltm), abs(rtm)). TRUE = (ltm, rtm). o Change the page names and collate order Changes in version 1.6-2 o new functions: checkcoefk(), replaceNA() o new function: ckiener7() deals with p vector and coefk matrix (recycling) o dimdim(NULL) returns now c(0,0) (before c(1,0)) o checkquantiles() rewritten with a stop by default except if STOP = FALSE o parallel::makeCluster() now requires package {methods} Changes in version 1.6-1 o fatreturns, logret and elevate transferred from FatTailsRplot to FatTailsR Changes in version 1.6-0 (2016-02-20) o Codename: Caucus in Nevada o Bug when lower.tail was FALSE. New version is such that: ltmkiener?(p, ..., lower.tail=FALSE) == rtmkiener?(p, ..., lower.tail=TRUE) rtmkiener?(p, ..., lower.tail=FALSE) == ltmkiener?(p, ..., lower.tail=TRUE) eskiener?() =ltmkiener? for left tail, =rtmkiener? for right tail and always positive Changes in version 1.5-8 o fitkienerX, paramkienerX, paramkienerX5, paramkienerX7 Recursive approach: accept list of vectors, matrix, timeSeries, xts, zoo... Changes in version 1.5-6 o R updated from R-3.1.0 to R-3.2.3 o Expected Shortfall added to {regkienerLX} (dfrESkPk and dfrESkLk) Changes in version 1.5-3 o Improved parallelization of arrays in {fitkienerX} and {paramkienerX57} Changes in version 1.5-0 (2015-11-08) o Codename: Rugby World cup - All Blacks win o Parallelization of arrays in {fitkienerX} and {paramkienerX57} o Bugs fixed: -- {fitkienerX} returned NA in columns {xmoments} ("m1x","sdx","skx","kex") -- {dimdim} returned wrong values with data.frame Changes in version 1.4-1 (2015-10-26) o Codename: Rugby World cup o A major update !! o 3 months of intensive work for better performance and higher flexibility o Compatible with commercial package FatTailsRplot (several advanced plots) -- Plots can be seen in 2014 and 2015 presentations at R/Rmetrics Conference -- Contact me for a demo NEW FUNCTIONS o Functions {fitkienerX} and {paramkienerX}: -- accomodate vectors, matrix, data.frame, arrays and lists -- replace {fitkienerLX} and {estimkienerX} (1.2-0) which are suppressed (we apologize to early users who will have to rewrite their code) The new code is much better. -- same output than {regkienerLX}$fitk -- some parameters renamed ("g.01" changed to "dg.01") -- smallest dataset must have more than 15 valid quantiles. -- return a huge vector {fitk} with quantiles, moments, VaR, ES, etc... o Predefined {exfit0},..,{exfit7} to subset this vector {fitk} o {pprobs0},..,{pprobs9} as preset values of probabilities to be combined with o {getnamesk} to generate parameter names o p1 replaces 1-p1 in elevenprobs (14 august 2015) o {dimdim}, {dimdim1}: alternate version to {dim} (03 August 2015) o {roundcoefk} to simplify parameter rounding (+ preset values) (1 August 2015, 1.2-6) o {ltmkiener1234}, {rtmkiener1234} {eskiener1234} Left tail mean and Right tail mean as signed (+/-) Expected Shortfall for left and right tails (18 September 2015) o {hkiener1234} generate parameter h.01,..,h.99 (ES/ES_logis) (18 October 2015) o {ExtractData} as a wrapper of tData, xData, zData, dfData o Colnames with upper cases in datasets df,t,x,zData (14 August 2015) o {kmoments}, {xmoments} examples rewritten Changes in version 1.2-3 (2015-07-08) o Code 1.2-0 made compliant with R-devel (next R 3.3-0). Changes in version 1.2-0 (2015-06-18) o Codename: Appel du 18 juin o Title changed to: Kiener Distributions and Fat Tails in Finance. o Raw and central moments. o regkienerLX has now a @exportClass(clregk). o fitkienerLX: same algorithm than regkienerLX but displays a flat table. o estimkienerX: new algorithm based on 5, 7, 9 or 11 quantiles. o {checkquantiles} o {pk2pk} conversion function o {ckiener1234} and parameter {c.01},..,{c.99} Changes in version 1.0-3 (2014-07-14) o Codename : Liberte, Egalite, Fraternite o Initial release. o Title: Power Hyperbolic Functions and Kiener Distributions. o Developped with R version R-3.1.0