------------------------------------------------------------------------
r522 | mathieu_ribatet | 2010-01-14 11:48:16 +0100 (jeu. 14 janv. 2010) | 2 lignes

Fix a bug in the geomgaussstderr (marginal part)

------------------------------------------------------------------------
r521 | mathieu_ribatet | 2010-01-14 11:47:32 +0100 (jeu. 14 janv. 2010) | 2 lignes

Slight optimizations of the C code

------------------------------------------------------------------------
r520 | mathieu_ribatet | 2010-01-13 23:40:16 +0100 (mer. 13 janv. 2010) | 2 lignes

Fix a bug in the computation of standard errors for geomgauss

------------------------------------------------------------------------
r519 | mathieu_ribatet | 2010-01-13 23:39:19 +0100 (mer. 13 janv. 2010) | 2 lignes

Minor typo in the documentation of map and condmap

------------------------------------------------------------------------
r518 | mathieu_ribatet | 2010-01-13 23:38:23 +0100 (mer. 13 janv. 2010) | 3 lignes

gev2frech uses the rank instead of ppoints for the empirical
distribution - better for the F-madogram

------------------------------------------------------------------------
r517 | mathieu_ribatet | 2010-01-13 23:37:25 +0100 (mer. 13 janv. 2010) | 2 lignes

Add data sets and function to plot a map of Switzerland

------------------------------------------------------------------------
r516 | mathieu_ribatet | 2010-01-13 23:36:59 +0100 (mer. 13 janv. 2010) | 2 lignes

Add data sets and function to plot a map of Switzerland

------------------------------------------------------------------------
r515 | mathieu_ribatet | 2010-01-13 23:35:43 +0100 (mer. 13 janv. 2010) | 2 lignes

Forgot to put the names gradient in the std err computations

------------------------------------------------------------------------
r514 | mathieu_ribatet | 2010-01-13 23:34:52 +0100 (mer. 13 janv. 2010) | 2 lignes

Remove the dots argument in uniroot in condmap

------------------------------------------------------------------------
r513 | mathieu_ribatet | 2010-01-13 23:33:46 +0100 (mer. 13 janv. 2010) | 2 lignes

Define new values for the Brown-Resnick processes

------------------------------------------------------------------------
r512 | mathieu_ribatet | 2010-01-06 16:10:44 +0100 (mer. 06 janv. 2010) | 2 lignes

Now we use 'crossprod' which is faster when std.err.type = 'grad'

------------------------------------------------------------------------
r511 | mathieu_ribatet | 2009-12-29 18:40:01 +0100 (mar. 29 déc. 2009) | 2 lignes

Use the standard given by R for the NEWS file

------------------------------------------------------------------------
r510 | mathieu_ribatet | 2009-12-29 18:10:05 +0100 (mar. 29 déc. 2009) | 2 lignes

Add a NEWS file

------------------------------------------------------------------------
r509 | mathieu_ribatet | 2009-12-29 17:51:55 +0100 (mar. 29 déc. 2009) | 2 lignes

Update

------------------------------------------------------------------------
r508 | mathieu_ribatet | 2009-12-29 17:51:07 +0100 (mar. 29 déc. 2009) | 2 lignes

Update the ChangeLog

------------------------------------------------------------------------
r507 | mathieu_ribatet | 2009-12-29 17:37:33 +0100 (mar. 29 déc. 2009) | 2 lignes

Fix a typo

------------------------------------------------------------------------
r506 | mathieu_ribatet | 2009-12-29 17:21:50 +0100 (mar. 29 déc. 2009) | 2 lignes

Remove the Figures folder - useless

------------------------------------------------------------------------
r505 | mathieu_ribatet | 2009-12-29 17:19:30 +0100 (mar. 29 déc. 2009) | 2 lignes

The example is now excuted and nicer

------------------------------------------------------------------------
r504 | mathieu_ribatet | 2009-12-29 17:00:41 +0100 (mar. 29 déc. 2009) | 2 lignes

Add an invisible frech2gev function which is faster but less flexible

------------------------------------------------------------------------
r503 | mathieu_ribatet | 2009-12-29 16:59:03 +0100 (mar. 29 déc. 2009) | 3 lignes

map and condmap can now deal with covariates through the 'covariates'
argument

------------------------------------------------------------------------
r502 | mathieu_ribatet | 2009-12-29 16:57:36 +0100 (mar. 29 déc. 2009) | 2 lignes

Some code optimizations

------------------------------------------------------------------------
r501 | mathieu_ribatet | 2009-12-29 16:56:42 +0100 (mar. 29 déc. 2009) | 2 lignes

Allow the use of weights in the pairwise likelihood

------------------------------------------------------------------------
r500 | mathieu_ribatet | 2009-12-17 12:03:37 +0100 (jeu. 17 déc. 2009) | 2 lignes

Improve the examples

------------------------------------------------------------------------
r499 | mathieu_ribatet | 2009-12-17 12:03:12 +0100 (jeu. 17 déc. 2009) | 2 lignes

Improve how the x and y axis limits are computed

------------------------------------------------------------------------
r498 | mathieu_ribatet | 2009-12-17 11:24:24 +0100 (jeu. 17 déc. 2009) | 2 lignes

Add an option "std.err" to compute or/not the standard errors

------------------------------------------------------------------------
r497 | mathieu_ribatet | 2009-12-17 11:23:56 +0100 (jeu. 17 déc. 2009) | 2 lignes

Rewrite the map function to allow the use of covariates if necessary

------------------------------------------------------------------------
r496 | mathieu_ribatet | 2009-12-16 20:54:46 +0100 (mer. 16 déc. 2009) | 2 lignes

Add the display of the TIC for object of class spatgev

------------------------------------------------------------------------
r495 | mathieu_ribatet | 2009-12-16 20:53:54 +0100 (mer. 16 déc. 2009) | 2 lignes

Comment fitnsmaxstab - not time to implement it

------------------------------------------------------------------------
r494 | mathieu_ribatet | 2009-12-16 20:53:30 +0100 (mer. 16 déc. 2009) | 2 lignes

Remove rgl

------------------------------------------------------------------------
r493 | mathieu_ribatet | 2009-12-16 20:51:32 +0100 (mer. 16 déc. 2009) | 2 lignes

Removing an unused declared variable

------------------------------------------------------------------------
r491 | mathieu_ribatet | 2009-12-16 12:02:53 +0100 (mer. 16 déc. 2009) | 2 lignes

No warning message if fitspatgev failed to converge

------------------------------------------------------------------------
r490 | mathieu_ribatet | 2009-12-16 12:00:40 +0100 (mer. 16 déc. 2009) | 3 lignes

Move nsgeomgauss.R to the development folder - no implementation
planned

------------------------------------------------------------------------
r489 | mathieu_ribatet | 2009-12-16 11:27:15 +0100 (mer. 16 déc. 2009) | 2 lignes

All the (unecessary) gradient functions were removed

------------------------------------------------------------------------
r488 | mathieu_ribatet | 2009-12-16 11:26:21 +0100 (mer. 16 déc. 2009) | 3 lignes

The fitspatgev function has now a "corr" argument to be consistent
with the fitmaxstab function.

------------------------------------------------------------------------
r487 | mathieu_ribatet | 2009-12-16 11:25:42 +0100 (mer. 16 déc. 2009) | 2 lignes

Improving the code

------------------------------------------------------------------------
r486 | mathieu_ribatet | 2009-12-16 11:24:55 +0100 (mer. 16 déc. 2009) | 3 lignes

Remove the use of the rgl package (warnings when building the package)
and passing marge = "emp" as the default

------------------------------------------------------------------------
r485 | mathieu_ribatet | 2009-12-16 11:24:05 +0100 (mer. 16 déc. 2009) | 2 lignes

Improve the way how the asymptotic covariance matrices are computed

------------------------------------------------------------------------
r484 | mathieu_ribatet | 2009-12-16 11:22:21 +0100 (mer. 16 déc. 2009) | 3 lignes

What I previously called jacobian was misleading and now I used
var.score standing for "variance of the (composite) score"

------------------------------------------------------------------------
r483 | mathieu_ribatet | 2009-12-16 11:20:34 +0100 (mer. 16 déc. 2009) | 3 lignes

Removing the gradient files: now useless because of standardErrors
files

------------------------------------------------------------------------
r482 | mathieu_ribatet | 2009-12-15 23:46:05 +0100 (mar. 15 déc. 2009) | 2 lignes

Update the ChangLog

------------------------------------------------------------------------
r481 | mathieu_ribatet | 2009-12-15 23:39:40 +0100 (mar. 15 déc. 2009) | 2 lignes

Update the version number

------------------------------------------------------------------------
r480 | mathieu_ribatet | 2009-12-15 23:36:09 +0100 (mar. 15 déc. 2009) | 2 lignes

Small editing in view of a new release

------------------------------------------------------------------------
r479 | mathieu_ribatet | 2009-12-15 23:31:26 +0100 (mar. 15 déc. 2009) | 2 lignes

Fix a typo

------------------------------------------------------------------------
r478 | mathieu_ribatet | 2009-12-15 23:28:41 +0100 (mar. 15 déc. 2009) | 2 lignes

Forgot that by default lmadogram uses an rgl output

------------------------------------------------------------------------
r477 | mathieu_ribatet | 2009-12-15 23:27:00 +0100 (mar. 15 déc. 2009) | 3 lignes

Revert to a former version and reduce the grid sizes to get less heavy
pdf output

------------------------------------------------------------------------
r476 | mathieu_ribatet | 2009-12-15 23:23:35 +0100 (mar. 15 déc. 2009) | 2 lignes

Update ref

------------------------------------------------------------------------
r475 | mathieu_ribatet | 2009-12-15 23:22:18 +0100 (mar. 15 déc. 2009) | 2 lignes

New way to compute the asymptotic covariance matrix - header

------------------------------------------------------------------------
r474 | mathieu_ribatet | 2009-12-15 23:09:09 +0100 (mar. 15 déc. 2009) | 2 lignes

Fix a bug in smith3d

------------------------------------------------------------------------
r473 | mathieu_ribatet | 2009-12-15 23:07:43 +0100 (mar. 15 déc. 2009) | 2 lignes

New way to compute the asymptotic covariance matrix

------------------------------------------------------------------------
r472 | mathieu_ribatet | 2009-12-15 17:21:41 +0100 (mar. 15 déc. 2009) | 2 lignes

Use marge = "emp" by defaults - faster

------------------------------------------------------------------------
r471 | mathieu_ribatet | 2009-12-15 17:21:13 +0100 (mar. 15 déc. 2009) | 2 lignes

Small improvements

------------------------------------------------------------------------
r470 | mathieu_ribatet | 2009-12-15 17:20:48 +0100 (mar. 15 déc. 2009) | 2 lignes

Changes the methods as now Brown-Resnick process are available

------------------------------------------------------------------------
r469 | mathieu_ribatet | 2009-12-15 17:19:40 +0100 (mar. 15 déc. 2009) | 2 lignes

Changes the headers because of new functions

------------------------------------------------------------------------
r468 | mathieu_ribatet | 2009-12-15 17:19:18 +0100 (mar. 15 déc. 2009) | 2 lignes

Add the generalized Cauchy correlation function

------------------------------------------------------------------------
r467 | mathieu_ribatet | 2009-12-15 17:18:56 +0100 (mar. 15 déc. 2009) | 2 lignes

Add the generalized Cauchy correlation function

------------------------------------------------------------------------
r466 | mathieu_ribatet | 2009-12-15 17:18:38 +0100 (mar. 15 déc. 2009) | 2 lignes

Add the generalized Cauchy correlation function

------------------------------------------------------------------------
r465 | mathieu_ribatet | 2009-12-15 17:18:15 +0100 (mar. 15 déc. 2009) | 3 lignes

Add the generalized Cauchy correlation function and Brown-Resnick
processes

------------------------------------------------------------------------
r464 | mathieu_ribatet | 2009-12-15 17:17:50 +0100 (mar. 15 déc. 2009) | 3 lignes

Add the generalized Cauchy correlation function and Brown-Resnick
processes

------------------------------------------------------------------------
r463 | mathieu_ribatet | 2009-12-15 17:17:08 +0100 (mar. 15 déc. 2009) | 4 lignes

This is a wrapper to the fitcovmat and fitcovariance function - to be
consistent with the fitmaxstab function. "ls" stands for least
squares.

------------------------------------------------------------------------
r462 | mathieu_ribatet | 2009-12-15 17:16:31 +0100 (mar. 15 déc. 2009) | 2 lignes

Add the Brown-Resnick processes

------------------------------------------------------------------------
r461 | mathieu_ribatet | 2009-12-15 17:16:16 +0100 (mar. 15 déc. 2009) | 3 lignes

Add the generalized Cauchy correlation function and Brown-Resnick
processes

------------------------------------------------------------------------
r460 | mathieu_ribatet | 2009-12-15 17:15:51 +0100 (mar. 15 déc. 2009) | 2 lignes

Use marge = "emp" by default and allow for non weighted least squares

------------------------------------------------------------------------
r459 | mathieu_ribatet | 2009-12-15 17:14:43 +0100 (mar. 15 déc. 2009) | 2 lignes

Add the generalized Cauchy correlation function

------------------------------------------------------------------------
r458 | mathieu_ribatet | 2009-12-15 17:14:27 +0100 (mar. 15 déc. 2009) | 2 lignes

Add the composite version of the BIC

------------------------------------------------------------------------
r457 | mathieu_ribatet | 2009-12-15 17:13:53 +0100 (mar. 15 déc. 2009) | 3 lignes

Add the generalized Cauchy correlation function and the computations
of starting values for Brown-Resnick processes

------------------------------------------------------------------------
r456 | mathieu_ribatet | 2009-12-15 17:13:13 +0100 (mar. 15 déc. 2009) | 2 lignes

Add the composite version of the BIC and remove duplicated code

------------------------------------------------------------------------
r455 | mathieu_ribatet | 2009-12-15 17:12:34 +0100 (mar. 15 déc. 2009) | 2 lignes

Add the generalized Cauchy correlation function

------------------------------------------------------------------------
r454 | mathieu_ribatet | 2009-12-15 17:11:49 +0100 (mar. 15 déc. 2009) | 2 lignes

Get standard errors for the predicted values if possible

------------------------------------------------------------------------
r453 | mathieu_ribatet | 2009-12-15 17:10:46 +0100 (mar. 15 déc. 2009) | 2 lignes

Now it works without any problem for one dimensional index set

------------------------------------------------------------------------
r452 | mathieu_ribatet | 2009-12-15 17:09:02 +0100 (mar. 15 déc. 2009) | 2 lignes

Add the generalized cauchy covariance function

------------------------------------------------------------------------
r451 | mathieu_ribatet | 2009-12-15 17:08:39 +0100 (mar. 15 déc. 2009) | 2 lignes

Add the brown-resnick processes

------------------------------------------------------------------------
r450 | mathieu_ribatet | 2009-10-31 11:46:32 +0100 (sam. 31 oct. 2009) | 2 lignes

Fix a bug when simulating the process on a grid: wrong iwork size

------------------------------------------------------------------------
r449 | mathieu_ribatet | 2009-10-31 11:46:11 +0100 (sam. 31 oct. 2009) | 2 lignes

Fix a bug when simulating the process on a grid

------------------------------------------------------------------------
r448 | mathieu_ribatet | 2009-10-31 11:44:56 +0100 (sam. 31 oct. 2009) | 2 lignes

Fix typos in the doc

------------------------------------------------------------------------
r447 | mathieu_ribatet | 2009-10-16 10:42:42 +0200 (ven. 16 oct. 2009) | 2 lignes

Update the version and ChangeLog

------------------------------------------------------------------------
r446 | mathieu_ribatet | 2009-10-16 10:37:02 +0200 (ven. 16 oct. 2009) | 2 lignes

Remove the inline attribute - not portable

------------------------------------------------------------------------
r445 | mathieu_ribatet | 2009-10-12 08:18:05 +0200 (lun. 12 oct. 2009) | 2 lignes

Minor change

------------------------------------------------------------------------
r444 | mathieu_ribatet | 2009-10-12 08:17:05 +0200 (lun. 12 oct. 2009) | 4 lignes

Pass the R_InLine function to the lplik??? function. When the
bivariate distribution is numerically degenerates, then it now returns
MINf instead of the computation of the univariate llik.

------------------------------------------------------------------------
r443 | mathieu_ribatet | 2009-10-12 08:14:37 +0200 (lun. 12 oct. 2009) | 2 lignes

Minor change

------------------------------------------------------------------------
r442 | mathieu_ribatet | 2009-10-12 08:14:17 +0200 (lun. 12 oct. 2009) | 2 lignes

Improvements in the variable definitions

------------------------------------------------------------------------
r441 | mathieu_ribatet | 2009-10-12 08:12:05 +0200 (lun. 12 oct. 2009) | 3 lignes

Minor changes to the documention because of complaints of the new Rd
parser

------------------------------------------------------------------------
r440 | mathieu_ribatet | 2009-10-12 08:11:06 +0200 (lun. 12 oct. 2009) | 2 lignes

Update the references

------------------------------------------------------------------------
r439 | mathieu_ribatet | 2009-08-24 14:32:22 +0200 (lun. 24 août 2009) | 2 lignes

Some small editing

------------------------------------------------------------------------
r438 | mathieu_ribatet | 2009-08-24 14:31:15 +0200 (lun. 24 août 2009) | 4 lignes

The argument control is now added to control several tuning parameters
for the simulation algo hence the prototype for the C functions are
updated

------------------------------------------------------------------------
r437 | mathieu_ribatet | 2009-08-24 14:29:28 +0200 (lun. 24 août 2009) | 2 lignes

Add the entry for the simulation of random fields

------------------------------------------------------------------------
r436 | mathieu_ribatet | 2009-08-24 14:29:00 +0200 (lun. 24 août 2009) | 3 lignes

The argument control is now added to control several tuning parameters
for the simulation algo

------------------------------------------------------------------------
r435 | mathieu_ribatet | 2009-08-24 14:25:56 +0200 (lun. 24 août 2009) | 2 lignes

If dist is a matrix it now returns a matrix too

------------------------------------------------------------------------
r434 | mathieu_ribatet | 2009-07-08 15:19:50 +0200 (mer. 08 juil. 2009) | 2 lignes

Add the simulation of the geometric gaussian process

------------------------------------------------------------------------
r433 | mathieu_ribatet | 2009-07-08 15:19:38 +0200 (mer. 08 juil. 2009) | 2 lignes

Fix a bug in the computation of the gradient - e.g. gwhitmat

------------------------------------------------------------------------
r432 | mathieu_ribatet | 2009-07-08 15:19:03 +0200 (mer. 08 juil. 2009) | 2 lignes

Fix a problem when using nlm with the Smith model

------------------------------------------------------------------------
r431 | mathieu_ribatet | 2009-07-08 15:18:40 +0200 (mer. 08 juil. 2009) | 2 lignes

Add the simulation of the geometric gaussian process

------------------------------------------------------------------------
r430 | mathieu_ribatet | 2009-07-08 15:17:29 +0200 (mer. 08 juil. 2009) | 2 lignes

Fix a typo

------------------------------------------------------------------------
r429 | mathieu_ribatet | 2009-06-30 10:28:13 +0200 (mar. 30 juin 2009) | 2 lignes

Fix a bug in the map function - shape parameter maps were not produced

------------------------------------------------------------------------
r428 | mathieu_ribatet | 2009-06-30 10:25:22 +0200 (mar. 30 juin 2009) | 2 lignes

Fix a typo

------------------------------------------------------------------------
r427 | mathieu_ribatet | 2009-05-27 16:11:41 +0200 (mer. 27 mai 2009) | 2 lignes

Initialize dist within this function to avoid initialization outside

------------------------------------------------------------------------
r426 | mathieu_ribatet | 2009-05-27 16:11:04 +0200 (mer. 27 mai 2009) | 2 lignes

Minor changes to avoid warnings from the compilers

------------------------------------------------------------------------
r425 | mathieu_ribatet | 2009-05-26 15:49:03 +0200 (mar. 26 mai 2009) | 2 lignes

Update changelog and release version

------------------------------------------------------------------------
r424 | mathieu_ribatet | 2009-05-26 15:48:02 +0200 (mar. 26 mai 2009) | 2 lignes

Minor change to Makevars

------------------------------------------------------------------------
r423 | mathieu_ribatet | 2009-05-26 15:39:29 +0200 (mar. 26 mai 2009) | 2 lignes

Cleaning C code

------------------------------------------------------------------------
r422 | mathieu_ribatet | 2009-05-26 15:39:21 +0200 (mar. 26 mai 2009) | 2 lignes

Cleaning C code

------------------------------------------------------------------------
r421 | mathieu_ribatet | 2009-05-26 15:33:32 +0200 (mar. 26 mai 2009) | 2 lignes

Update the vignette

------------------------------------------------------------------------
r420 | mathieu_ribatet | 2009-05-18 16:44:24 +0200 (lun. 18 mai 2009) | 3 lignes

The pairwise likelihood now uses univariate density in presence of
"numerical" complete dependence

------------------------------------------------------------------------
r419 | mathieu_ribatet | 2009-05-18 16:43:04 +0200 (lun. 18 mai 2009) | 2 lignes

Changes due to the implementation of new C code and the use of Lapack

------------------------------------------------------------------------
r418 | mathieu_ribatet | 2009-05-18 16:42:29 +0200 (lun. 18 mai 2009) | 2 lignes

Minor changes - boundary cases

------------------------------------------------------------------------
r417 | mathieu_ribatet | 2009-05-18 16:41:35 +0200 (lun. 18 mai 2009) | 3 lignes

The examples now use my functions instead of the ones provided by the
RandomFields package

------------------------------------------------------------------------
r416 | mathieu_ribatet | 2009-05-18 16:40:28 +0200 (lun. 18 mai 2009) | 2 lignes

Removed unnecessary file

------------------------------------------------------------------------
r415 | mathieu_ribatet | 2009-05-18 16:39:54 +0200 (lun. 18 mai 2009) | 3 lignes

New code related to the simulation of stationary max-stable random
fields

------------------------------------------------------------------------
r414 | mathieu_ribatet | 2009-05-18 16:38:43 +0200 (lun. 18 mai 2009) | 3 lignes

New code related to the simulation of stationary gaussian processes by
using the direct and the turning bands methods

------------------------------------------------------------------------
r413 | mathieu_ribatet | 2009-05-18 16:36:26 +0200 (lun. 18 mai 2009) | 3 lignes

Use the isotropic definition for the extremal coefficient if iso =
TRUE

------------------------------------------------------------------------
r412 | mathieu_ribatet | 2009-05-18 16:35:49 +0200 (lun. 18 mai 2009) | 3 lignes

Allow for the sill to be larger than 1 as now gaussian processes are
implemented

------------------------------------------------------------------------
r411 | mathieu_ribatet | 2009-05-18 16:35:07 +0200 (lun. 18 mai 2009) | 3 lignes

remove the definition of parscale in optim and avoid using Nelder-Mead
optimizer for one-dimensional problems

------------------------------------------------------------------------
r410 | mathieu_ribatet | 2009-05-05 23:39:13 +0200 (mar. 05 mai 2009) | 2 lignes

Extend the simulation from the Smith model

------------------------------------------------------------------------
r409 | mathieu_ribatet | 2009-05-05 23:37:23 +0200 (mar. 05 mai 2009) | 3 lignes

Remove the qr decomposition for the computation of the inverse of the
hessian matrix as it seems to be less robust

------------------------------------------------------------------------
r408 | mathieu_ribatet | 2009-05-05 23:35:46 +0200 (mar. 05 mai 2009) | 3 lignes

Add the iso as an argument of the output and compute the special
version of the extremal coefficient function when iso = TRUE

------------------------------------------------------------------------
r407 | mathieu_ribatet | 2009-04-30 17:04:57 +0200 (jeu. 30 avril 2009) | 3 lignes

A first version of a function that generates observations from a Smith
random field

------------------------------------------------------------------------
r406 | mathieu_ribatet | 2009-04-30 17:03:07 +0200 (jeu. 30 avril 2009) | 2 lignes

Minor changes

------------------------------------------------------------------------
r405 | mathieu_ribatet | 2009-04-30 17:02:49 +0200 (jeu. 30 avril 2009) | 2 lignes

Add the control in optim as an argument

------------------------------------------------------------------------
r404 | mathieu_ribatet | 2009-04-30 17:02:14 +0200 (jeu. 30 avril 2009) | 2 lignes

Minor changes

------------------------------------------------------------------------
r403 | mathieu_ribatet | 2009-04-30 17:01:50 +0200 (jeu. 30 avril 2009) | 2 lignes

Some minor changes: control in optim, ...

------------------------------------------------------------------------
r402 | mathieu_ribatet | 2009-04-30 17:01:18 +0200 (jeu. 30 avril 2009) | 3 lignes

If iso = TRUE in the smith model the extremal coefficient function is
then one-dimensional

------------------------------------------------------------------------
r401 | mathieu_ribatet | 2009-04-30 17:00:19 +0200 (jeu. 30 avril 2009) | 3 lignes

The madogram plots now accept a fitted model as an argument to compare
pairwise estimates and model based ones

------------------------------------------------------------------------
r400 | mathieu_ribatet | 2009-04-30 16:58:54 +0200 (jeu. 30 avril 2009) | 2 lignes

Implementation of the Bessel covariance function

------------------------------------------------------------------------
r399 | mathieu_ribatet | 2009-04-20 11:31:24 +0200 (lun. 20 avril 2009) | 2 lignes

Update the changelog

------------------------------------------------------------------------
r398 | mathieu_ribatet | 2009-04-17 17:48:09 +0200 (ven. 17 avril 2009) | 3 lignes

Fix a bug in the computation of the gradient for smith3d. Now it's
correct!

------------------------------------------------------------------------
r397 | mathieu_ribatet | 2009-04-17 17:29:18 +0200 (ven. 17 avril 2009) | 2 lignes

Minor optimization

------------------------------------------------------------------------
r396 | mathieu_ribatet | 2009-04-17 17:28:56 +0200 (ven. 17 avril 2009) | 2 lignes

Try to optimize the computation of all the covariance

------------------------------------------------------------------------
r395 | mathieu_ribatet | 2009-04-10 15:37:53 +0200 (ven. 10 avril 2009) | 2 lignes

Update the ChangeLog

------------------------------------------------------------------------
r394 | mathieu_ribatet | 2009-04-10 15:35:37 +0200 (ven. 10 avril 2009) | 2 lignes

Update the version number and the ChangeLog file

------------------------------------------------------------------------
r393 | mathieu_ribatet | 2009-04-10 12:06:58 +0200 (ven. 10 avril 2009) | 2 lignes

Minor changes

------------------------------------------------------------------------
r392 | mathieu_ribatet | 2009-04-10 12:06:36 +0200 (ven. 10 avril 2009) | 2 lignes

Add the computation of the gradient for the geometric gaussian model

------------------------------------------------------------------------
r391 | mathieu_ribatet | 2009-04-09 16:18:07 +0200 (jeu. 09 avril 2009) | 2 lignes

Minor optimizations

------------------------------------------------------------------------
r390 | mathieu_ribatet | 2009-04-09 16:17:22 +0200 (jeu. 09 avril 2009) | 2 lignes

Minor optimizations - and fix a bug in Schlather's gradient

------------------------------------------------------------------------
r389 | mathieu_ribatet | 2009-04-09 16:16:04 +0200 (jeu. 09 avril 2009) | 2 lignes

Minor optimizations

------------------------------------------------------------------------
r388 | mathieu_ribatet | 2009-04-07 17:45:18 +0200 (mar. 07 avril 2009) | 2 lignes

Some minor optimizations

------------------------------------------------------------------------
r387 | mathieu_ribatet | 2009-04-07 17:44:53 +0200 (mar. 07 avril 2009) | 2 lignes

Bug : replace '|' by '||' i.e. logical operator

------------------------------------------------------------------------
r386 | mathieu_ribatet | 2009-04-07 17:44:40 +0200 (mar. 07 avril 2009) | 2 lignes

Minor change

------------------------------------------------------------------------
r385 | mathieu_ribatet | 2009-04-07 17:43:00 +0200 (mar. 07 avril 2009) | 2 lignes

Bug : replace '|' by '||' i.e. logical operator

------------------------------------------------------------------------
r384 | mathieu_ribatet | 2009-04-07 17:42:48 +0200 (mar. 07 avril 2009) | 2 lignes

Bug : replace '|' by '||' i.e. logical operator

------------------------------------------------------------------------
r383 | mathieu_ribatet | 2009-04-07 17:42:33 +0200 (mar. 07 avril 2009) | 2 lignes

Bug : replace '|' by '||' i.e. logical operator

------------------------------------------------------------------------
r382 | mathieu_ribatet | 2009-04-07 17:42:19 +0200 (mar. 07 avril 2009) | 2 lignes

Optimize the computation of the pairwise likelihood

------------------------------------------------------------------------
r381 | mathieu_ribatet | 2009-04-07 17:42:00 +0200 (mar. 07 avril 2009) | 2 lignes

Optimize the computation of the pairwise likelihood

------------------------------------------------------------------------
r380 | mathieu_ribatet | 2009-04-07 17:41:28 +0200 (mar. 07 avril 2009) | 2 lignes

Bug : replace '|' by '||' i.e. logical operator

------------------------------------------------------------------------
r379 | mathieu_ribatet | 2009-04-07 17:41:16 +0200 (mar. 07 avril 2009) | 2 lignes

Optimize the computation of the gradient and fix some errors

------------------------------------------------------------------------
r378 | mathieu_ribatet | 2009-04-07 17:40:47 +0200 (mar. 07 avril 2009) | 2 lignes

Bug : replace '|' by '||' i.e. logical operator

------------------------------------------------------------------------
r377 | mathieu_ribatet | 2009-04-07 17:40:17 +0200 (mar. 07 avril 2009) | 2 lignes

Optmize the computation of weighted sum of squares

------------------------------------------------------------------------
r376 | mathieu_ribatet | 2009-04-07 17:39:45 +0200 (mar. 07 avril 2009) | 2 lignes

Optmize the computation of pairwise covariance

------------------------------------------------------------------------
r375 | mathieu_ribatet | 2009-04-07 17:39:00 +0200 (mar. 07 avril 2009) | 2 lignes

Minor change

------------------------------------------------------------------------
r374 | mathieu_ribatet | 2009-04-07 17:38:41 +0200 (mar. 07 avril 2009) | 2 lignes

Now fitspatgev uses all the optimizers as in fitmaxstab

------------------------------------------------------------------------
r373 | mathieu_ribatet | 2009-04-07 17:37:55 +0200 (mar. 07 avril 2009) | 2 lignes

Add argument jacobian = TRUE for consistency with other functions

------------------------------------------------------------------------
r372 | mathieu_ribatet | 2009-04-07 17:37:32 +0200 (mar. 07 avril 2009) | 2 lignes

Change the default starting values

------------------------------------------------------------------------
r371 | mathieu_ribatet | 2009-02-19 14:05:13 +0100 (jeu. 19 févr. 2009) | 2 lignes

Add references (Mehdi and Simone)

------------------------------------------------------------------------
r369 | mathieu_ribatet | 2009-02-17 14:46:36 +0100 (mar. 17 févr. 2009) | 2 lignes

Speed up the code

------------------------------------------------------------------------
r368 | mathieu_ribatet | 2009-02-17 14:45:55 +0100 (mar. 17 févr. 2009) | 2 lignes

Update the version number

------------------------------------------------------------------------
r367 | mathieu_ribatet | 2009-02-17 14:45:14 +0100 (mar. 17 févr. 2009) | 2 lignes

Update the vignette

------------------------------------------------------------------------
r366 | mathieu_ribatet | 2009-02-07 22:19:36 +0100 (sam. 07 févr. 2009) | 2 lignes

Update the ChangeLog and the DESCRIPTION file for the version 1.1-0

------------------------------------------------------------------------
r365 | mathieu_ribatet | 2009-02-07 22:10:28 +0100 (sam. 07 févr. 2009) | 2 lignes

Fix an error in the definition of a elliptical covariance function

------------------------------------------------------------------------
r364 | mathieu_ribatet | 2009-02-06 00:43:25 +0100 (ven. 06 févr. 2009) | 2 lignes

Update the vignette

------------------------------------------------------------------------
r363 | mathieu_ribatet | 2009-02-06 00:42:54 +0100 (ven. 06 févr. 2009) | 2 lignes

Fig a bug when held fixing parameter and defining starting values

------------------------------------------------------------------------
r361 | mathieu_ribatet | 2009-02-05 16:03:42 +0100 (jeu. 05 févr. 2009) | 2 lignes

Update the vignette

------------------------------------------------------------------------
r360 | mathieu_ribatet | 2009-02-05 16:03:18 +0100 (jeu. 05 févr. 2009) | 2 lignes

Redifine the GCV score so that it has the same scale as CV

------------------------------------------------------------------------
r356 | mathieu_ribatet | 2009-02-01 00:40:01 +0100 (dim. 01 févr. 2009) | 2 lignes

Found a more clever way to deal with 1-dimensional coordinates

------------------------------------------------------------------------
r355 | mathieu_ribatet | 2009-02-01 00:33:35 +0100 (dim. 01 févr. 2009) | 3 lignes

The function distance wasn't working if the coordinates where scalar,
fixed!

------------------------------------------------------------------------
r354 | mathieu_ribatet | 2009-01-31 00:52:44 +0100 (sam. 31 janv. 2009) | 3 lignes

Set weights that are really small (i.e. <= 1e-4) to more feasible
values

------------------------------------------------------------------------
r353 | mathieu_ribatet | 2009-01-18 18:36:21 +0100 (dim. 18 janv. 2009) | 2 lignes

The dots arguments was inverting with the start one

------------------------------------------------------------------------
r352 | mathieu_ribatet | 2009-01-18 18:35:48 +0100 (dim. 18 janv. 2009) | 2 lignes

Remove draft sections from the vignette

------------------------------------------------------------------------
r351 | mathieu_ribatet | 2009-01-18 18:35:17 +0100 (dim. 18 janv. 2009) | 2 lignes

Use different function names to avoid warning messages when compiling

------------------------------------------------------------------------
r350 | mathieu_ribatet | 2009-01-18 16:10:56 +0100 (dim. 18 janv. 2009) | 2 lignes

Improve how starting values are defined

------------------------------------------------------------------------
r349 | mathieu_ribatet | 2009-01-18 16:10:22 +0100 (dim. 18 janv. 2009) | 2 lignes

Improve how starting values are defined

------------------------------------------------------------------------
r348 | mathieu_ribatet | 2009-01-18 16:09:25 +0100 (dim. 18 janv. 2009) | 2 lignes

Improve how starting values are defined

------------------------------------------------------------------------
r347 | mathieu_ribatet | 2009-01-16 18:35:44 +0100 (ven. 16 janv. 2009) | 3 lignes

Allow the use of the independent Schlather and geometric Gaussian
models for the least squares fitting procedure

------------------------------------------------------------------------
r346 | mathieu_ribatet | 2009-01-15 19:31:56 +0100 (jeu. 15 janv. 2009) | 2 lignes

Fix a bug in defining the starting values for geomgaussfull

------------------------------------------------------------------------
r345 | mathieu_ribatet | 2009-01-15 00:19:38 +0100 (jeu. 15 janv. 2009) | 2 lignes

Minor changes

------------------------------------------------------------------------
r344 | mathieu_ribatet | 2009-01-14 14:24:22 +0100 (mer. 14 janv. 2009) | 2 lignes

Add some proofs

------------------------------------------------------------------------
r343 | mathieu_ribatet | 2009-01-14 14:23:57 +0100 (mer. 14 janv. 2009) | 3 lignes

Fix a bug. I omit to specify the covariance model when computing
appropriate starting values for the Schlather model

------------------------------------------------------------------------
r342 | mathieu_ribatet | 2009-01-14 14:23:06 +0100 (mer. 14 janv. 2009) | 2 lignes

We now use the corrected lambda-madogram - thanks to D. Cooley

------------------------------------------------------------------------
r341 | mathieu_ribatet | 2009-01-12 10:06:58 +0100 (lun. 12 janv. 2009) | 2 lignes

Extend the response surfaces section - not finished

------------------------------------------------------------------------
r340 | mathieu_ribatet | 2009-01-12 10:03:00 +0100 (lun. 12 janv. 2009) | 3 lignes

Add proof for the derivation of the Smith and geometric Gaussian
models

------------------------------------------------------------------------
r339 | mathieu_ribatet | 2009-01-06 16:06:12 +0100 (mar. 06 janv. 2009) | 2 lignes

Fix a wrong argument in the example

------------------------------------------------------------------------
r338 | mathieu_ribatet | 2009-01-03 02:58:06 +0100 (sam. 03 janv. 2009) | 3 lignes

Improve (a lot!) the package vignette. It's for from being finished
though...

------------------------------------------------------------------------
r337 | mathieu_ribatet | 2009-01-03 02:56:57 +0100 (sam. 03 janv. 2009) | 2 lignes

Minor changes

------------------------------------------------------------------------
r336 | mathieu_ribatet | 2009-01-02 18:43:46 +0100 (ven. 02 janv. 2009) | 2 lignes

Update reference list

------------------------------------------------------------------------
r335 | mathieu_ribatet | 2009-01-02 18:43:23 +0100 (ven. 02 janv. 2009) | 2 lignes

Move the qqextcoeff function to the modelChecking.R file

------------------------------------------------------------------------
r334 | mathieu_ribatet | 2009-01-02 18:42:56 +0100 (ven. 02 janv. 2009) | 2 lignes

Minor changes

------------------------------------------------------------------------
r333 | mathieu_ribatet | 2009-01-02 18:42:36 +0100 (ven. 02 janv. 2009) | 3 lignes

Now starting values for the GEV margins are defined by using the
fitspatgev function

------------------------------------------------------------------------
r332 | mathieu_ribatet | 2009-01-02 18:42:05 +0100 (ven. 02 janv. 2009) | 2 lignes

Minor changes

------------------------------------------------------------------------
r331 | mathieu_ribatet | 2009-01-02 18:41:38 +0100 (ven. 02 janv. 2009) | 3 lignes

Pass defaults values for parscale and change the initialization for
starting values

------------------------------------------------------------------------
r330 | mathieu_ribatet | 2009-01-02 18:40:29 +0100 (ven. 02 janv. 2009) | 2 lignes

Moving the madogram function to the madogram.R file

------------------------------------------------------------------------
r329 | mathieu_ribatet | 2009-01-02 18:39:59 +0100 (ven. 02 janv. 2009) | 2 lignes

If xlim is supplied the function doesn't compute appropriate x.range

------------------------------------------------------------------------
r328 | mathieu_ribatet | 2009-01-02 18:38:34 +0100 (ven. 02 janv. 2009) | 2 lignes

The header for the madogram-based approaches

------------------------------------------------------------------------
r327 | mathieu_ribatet | 2009-01-02 18:38:17 +0100 (ven. 02 janv. 2009) | 3 lignes

Optimize the code: now fitting a max-stable process by maximizing the
pairwise-likelihood is 40% times faster

------------------------------------------------------------------------
r326 | mathieu_ribatet | 2009-01-02 18:36:29 +0100 (ven. 02 janv. 2009) | 2 lignes

Update the documentation files

------------------------------------------------------------------------
r325 | mathieu_ribatet | 2009-01-02 18:35:33 +0100 (ven. 02 janv. 2009) | 2 lignes

Add figures for the geometric gaussian process - 2 simulations

------------------------------------------------------------------------
r324 | mathieu_ribatet | 2009-01-02 18:34:42 +0100 (ven. 02 janv. 2009) | 3 lignes

Add the madogram-based approaches, model checking and univariate EVD
functions

------------------------------------------------------------------------
r323 | mathieu_ribatet | 2008-12-19 16:54:56 +0100 (ven. 19 déc. 2008) | 2 lignes

Update the package Vignette

------------------------------------------------------------------------
r322 | mathieu_ribatet | 2008-12-19 11:03:11 +0100 (ven. 19 déc. 2008) | 2 lignes

Updating the main man page.

------------------------------------------------------------------------
r321 | mathieu_ribatet | 2008-12-19 11:00:10 +0100 (ven. 19 déc. 2008) | 2 lignes

Updating the vignette.

------------------------------------------------------------------------
r320 | mathieu_ribatet | 2008-12-18 09:35:47 +0100 (jeu. 18 déc. 2008) | 2 lignes

Wrong link to the profile2d.maxstab function

------------------------------------------------------------------------
r319 | mathieu_ribatet | 2008-12-17 22:18:54 +0100 (mer. 17 déc. 2008) | 2 lignes

Correct English and grammar thanks to Anthony

------------------------------------------------------------------------
r318 | mathieu_ribatet | 2008-12-17 21:22:32 +0100 (mer. 17 déc. 2008) | 3 lignes

Test if we have 0 weights and set a really huge value for non feasible
regions

------------------------------------------------------------------------
r317 | mathieu_ribatet | 2008-12-17 21:21:45 +0100 (mer. 17 déc. 2008) | 2 lignes

Changes due to the new way I deal with non feasible region

------------------------------------------------------------------------
r316 | mathieu_ribatet | 2008-12-16 19:49:19 +0100 (mar. 16 déc. 2008) | 2 lignes

Improve this functions

------------------------------------------------------------------------
r315 | mathieu_ribatet | 2008-12-16 19:48:43 +0100 (mar. 16 déc. 2008) | 3 lignes

The sill isn't fix for the starting values as it may cause troubles
when fixing it for the model definition

------------------------------------------------------------------------
r314 | mathieu_ribatet | 2008-12-16 19:47:44 +0100 (mar. 16 déc. 2008) | 2 lignes

Minor change: adding the logLik value in the output

------------------------------------------------------------------------
r313 | mathieu_ribatet | 2008-12-16 19:47:05 +0100 (mar. 16 déc. 2008) | 3 lignes

I forgot to pass the option iso = iso when computing the variance of
the score

------------------------------------------------------------------------
r312 | mathieu_ribatet | 2008-12-16 19:44:57 +0100 (mar. 16 déc. 2008) | 3 lignes

Fix a bug in the computation of the gradient (only present when
reponse surfaces are used for the GEV parameters)

------------------------------------------------------------------------
r311 | mathieu_ribatet | 2008-12-16 19:41:04 +0100 (mar. 16 déc. 2008) | 2 lignes

Do not set pointwise estimate to 1 or 2 if they lie outsite of [1,2]

------------------------------------------------------------------------
r310 | mathieu_ribatet | 2008-12-16 19:40:01 +0100 (mar. 16 déc. 2008) | 3 lignes

Final change for unfeasible regions: finally setting MINF to the
log-likelihood for unfeasible region is the "best" way

------------------------------------------------------------------------
r309 | mathieu_ribatet | 2008-12-16 19:38:09 +0100 (mar. 16 déc. 2008) | 2 lignes

Update and refine the documentation files

------------------------------------------------------------------------
r308 | mathieu_ribatet | 2008-12-12 17:22:15 +0100 (ven. 12 déc. 2008) | 2 lignes

Improve the code for speed

------------------------------------------------------------------------
r307 | mathieu_ribatet | 2008-12-11 22:25:20 +0100 (jeu. 11 déc. 2008) | 2 lignes

Using the optim as optimizer instead of nlm - more robust

------------------------------------------------------------------------
r306 | mathieu_ribatet | 2008-12-11 16:09:52 +0100 (jeu. 11 déc. 2008) | 2 lignes

Adjust the starting values

------------------------------------------------------------------------
r305 | mathieu_ribatet | 2008-12-11 16:09:14 +0100 (jeu. 11 déc. 2008) | 3 lignes

Computes the sandwich estimate of the asymptotic covariance
matrix. TIC and ANOVA are thus available.

------------------------------------------------------------------------
r304 | mathieu_ribatet | 2008-12-10 21:21:38 +0100 (mer. 10 déc. 2008) | 3 lignes

Add a function to fit a "spatial" gev model e.g. GEV parameters are
modelled through response surfaces: doc files.

------------------------------------------------------------------------
r303 | mathieu_ribatet | 2008-12-10 16:56:31 +0100 (mer. 10 déc. 2008) | 3 lignes

Fix a bug in testing if the scale and shape paremeters were badly
defined.

------------------------------------------------------------------------
r302 | mathieu_ribatet | 2008-12-10 16:55:42 +0100 (mer. 10 déc. 2008) | 4 lignes

Add a function to fit a "spatial" gev model e.g. GEV parameters are
modelled through response surfaces. Related methods and predictions
are also implemented.

------------------------------------------------------------------------
r301 | mathieu_ribatet | 2008-12-09 17:17:41 +0100 (mar. 09 déc. 2008) | 3 lignes

Add the non-stationary geom. gauss. model with response surfaces for
the GEV parameters

------------------------------------------------------------------------
r300 | mathieu_ribatet | 2008-12-07 13:54:27 +0100 (dim. 07 déc. 2008) | 2 lignes

Minor change

------------------------------------------------------------------------
r299 | mathieu_ribatet | 2008-12-07 13:18:17 +0100 (dim. 07 déc. 2008) | 2 lignes

Optimizing the C code: improve conditioning tests

------------------------------------------------------------------------
r298 | mathieu_ribatet | 2008-12-07 12:34:04 +0100 (dim. 07 déc. 2008) | 2 lignes

Proof-read the man pages due to a mail in the r-devel list

------------------------------------------------------------------------
r297 | mathieu_ribatet | 2008-12-06 16:34:56 +0100 (sam. 06 déc. 2008) | 2 lignes

Minor changes

------------------------------------------------------------------------
r296 | mathieu_ribatet | 2008-12-06 16:34:25 +0100 (sam. 06 déc. 2008) | 2 lignes

Fix a bug in the nsgeomgauss covariance function computations

------------------------------------------------------------------------
r295 | mathieu_ribatet | 2008-12-03 10:22:57 +0100 (mer. 03 déc. 2008) | 3 lignes

Do not use the same function name whatever the iso argument is to
avoid a warning message while building the package

------------------------------------------------------------------------
r294 | mathieu_ribatet | 2008-12-03 10:22:08 +0100 (mer. 03 déc. 2008) | 2 lignes

Masking the non-stationary geom. gauss. model for the moment

------------------------------------------------------------------------
r293 | mathieu_ribatet | 2008-12-03 10:21:07 +0100 (mer. 03 déc. 2008) | 2 lignes

There was an uninitialized variable

------------------------------------------------------------------------
r292 | mathieu_ribatet | 2008-12-03 09:33:30 +0100 (mer. 03 déc. 2008) | 2 lignes

Remove unused variable

------------------------------------------------------------------------
r291 | mathieu_ribatet | 2008-12-02 18:15:25 +0100 (mar. 02 déc. 2008) | 2 lignes

Update ChangeLog and DESCRIPTION for the new release

------------------------------------------------------------------------
r290 | mathieu_ribatet | 2008-12-02 18:02:52 +0100 (mar. 02 déc. 2008) | 2 lignes

Add a function that performs a QQ-plot for the extremal coefficient

------------------------------------------------------------------------
r289 | mathieu_ribatet | 2008-12-02 16:58:47 +0100 (mar. 02 déc. 2008) | 2 lignes

Change the way I deal with non-feasible regions

------------------------------------------------------------------------
r288 | mathieu_ribatet | 2008-12-02 16:58:08 +0100 (mar. 02 déc. 2008) | 2 lignes

Mask the non-stationary geom. gauss. model

------------------------------------------------------------------------
r287 | mathieu_ribatet | 2008-12-02 16:57:45 +0100 (mar. 02 déc. 2008) | 2 lignes

Implementation of an isotropic smith's model

------------------------------------------------------------------------
r286 | mathieu_ribatet | 2008-12-02 11:53:33 +0100 (mar. 02 déc. 2008) | 3 lignes

Change the way starting values are defined - i.e. fixing the sill
Change the way we deal with non feasible regions

------------------------------------------------------------------------
r285 | mathieu_ribatet | 2008-11-28 13:12:45 +0100 (ven. 28 nov. 2008) | 2 lignes

Add the geometric gaussian model - non stationary

------------------------------------------------------------------------
r284 | mathieu_ribatet | 2008-11-28 13:11:56 +0100 (ven. 28 nov. 2008) | 1 ligne


------------------------------------------------------------------------
r283 | mathieu_ribatet | 2008-11-27 15:42:04 +0100 (jeu. 27 nov. 2008) | 4 lignes

Add comments within the code.
Tweak how I deal with the unfeasible region when the GEV parameters
are ill-defined

------------------------------------------------------------------------
r282 | mathieu_ribatet | 2008-11-19 14:58:55 +0100 (mer. 19 nov. 2008) | 2 lignes

There was a duplicate of function .start.schlatherind

------------------------------------------------------------------------
r281 | mathieu_ribatet | 2008-11-19 12:36:55 +0100 (mer. 19 nov. 2008) | 3 lignes

Fix a bug in geomCovariance, sigma2 was not set to the right value
when it tends to get large

------------------------------------------------------------------------
r280 | mathieu_ribatet | 2008-11-19 12:29:28 +0100 (mer. 19 nov. 2008) | 2 lignes

Update some comments in the code

------------------------------------------------------------------------
r279 | mathieu_ribatet | 2008-11-19 12:20:46 +0100 (mer. 19 nov. 2008) | 2 lignes

Pass the RandomFields packages as a dependence

------------------------------------------------------------------------
r278 | mathieu_ribatet | 2008-11-19 12:16:53 +0100 (mer. 19 nov. 2008) | 2 lignes

Add an angle argument in madogram() to help in detecting anisotropy

------------------------------------------------------------------------
r277 | mathieu_ribatet | 2008-11-18 16:53:24 +0100 (mar. 18 nov. 2008) | 3 lignes

Fix a bug in schlatherinddsgnmat: the non feasible region was not
penalized for all cases

------------------------------------------------------------------------
r276 | mathieu_ribatet | 2008-11-17 13:57:30 +0100 (lun. 17 nov. 2008) | 3 lignes

Changing how to deal with smooth >= 2 for the powered exponential
covariance model

------------------------------------------------------------------------
r275 | mathieu_ribatet | 2008-10-31 17:45:39 +0100 (ven. 31 oct. 2008) | 3 lignes

Change the behaviour with non-feasible regions: seems to be more
robust now...

------------------------------------------------------------------------
r274 | mathieu_ribatet | 2008-10-28 14:44:28 +0100 (mar. 28 oct. 2008) | 3 lignes

Add an upper bound (i.e. 3) for the sigma2 parameter in the geometric
gaussian process

------------------------------------------------------------------------
r273 | mathieu_ribatet | 2008-10-24 09:54:01 +0200 (ven. 24 oct. 2008) | 3 lignes

Update the bounds for the covariance functions to avoid numerical
issues

------------------------------------------------------------------------
r271 | mathieu_ribatet | 2008-10-22 18:05:31 +0200 (mer. 22 oct. 2008) | 2 lignes

Add a link to the project web page

------------------------------------------------------------------------
r270 | mathieu_ribatet | 2008-10-22 14:23:42 +0200 (mer. 22 oct. 2008) | 3 lignes

Now the Gumbel distribution is considered whenever the shape parameter
is <= 1-6

------------------------------------------------------------------------
r269 | mathieu_ribatet | 2008-10-22 14:22:54 +0200 (mer. 22 oct. 2008) | 3 lignes

Change the madogram function. Now it is computed for each pairwise
distance - bins can be optionnaly passed though

------------------------------------------------------------------------
r268 | mathieu_ribatet | 2008-10-16 14:39:09 +0200 (jeu. 16 oct. 2008) | 3 lignes

Modifying the vignette and example to reduce the time to build the
package

------------------------------------------------------------------------
r264 | mathieu_ribatet | 2008-10-13 12:28:17 +0200 (lun. 13 oct. 2008) | 3 lignes

Cleaning the code: unused declared variables
Fix a mistake in computing the gradient dz2shape instead of dz2scale

------------------------------------------------------------------------
r262 | mathieu_ribatet | 2008-10-12 00:45:07 +0200 (dim. 12 oct. 2008) | 2 lignes

remove one copy of prior.Rd

------------------------------------------------------------------------
r261 | mathieu_ribatet | 2008-10-12 00:44:22 +0200 (dim. 12 oct. 2008) | 2 lignes

remove one copy of posterior.Rd

------------------------------------------------------------------------
r259 | mathieu_ribatet | 2008-10-09 10:53:37 +0200 (jeu. 09 oct. 2008) | 2 lignes

moving mcmc.R to the development folder

------------------------------------------------------------------------
r254 | mathieu_ribatet | 2008-10-07 15:45:11 +0200 (mar. 07 oct. 2008) | 3 lignes

Misplace conditional if resulting in cv = NA even when penalty = "gcv"
or "cv"

------------------------------------------------------------------------
r236 | mathieu_ribatet | 2008-10-07 00:12:55 +0200 (mar. 07 oct. 2008) | 2 lignes

Add the conditional quantile map and refine the documentation

------------------------------------------------------------------------
r231 | mathieu_ribatet | 2008-10-06 22:08:17 +0200 (lun. 06 oct. 2008) | 2 lignes

Moving the package to the /pkg folder

------------------------------------------------------------------------
r230 | mathieu_ribatet | 2008-10-06 22:07:30 +0200 (lun. 06 oct. 2008) | 2 lignes

Moving the package to the /pkg folder

------------------------------------------------------------------------
r229 | mathieu_ribatet | 2008-10-06 22:07:15 +0200 (lun. 06 oct. 2008) | 2 lignes

Moving the package to the /pkg folder

------------------------------------------------------------------------
r228 | mathieu_ribatet | 2008-10-06 22:07:05 +0200 (lun. 06 oct. 2008) | 2 lignes

Moving the package to the /pkg folder

------------------------------------------------------------------------
r227 | mathieu_ribatet | 2008-10-06 22:06:57 +0200 (lun. 06 oct. 2008) | 2 lignes

Moving the package to the /pkg folder

------------------------------------------------------------------------
r226 | mathieu_ribatet | 2008-10-06 22:06:45 +0200 (lun. 06 oct. 2008) | 2 lignes

Moving the package to the /pkg folder

------------------------------------------------------------------------
r221 | mathieu_ribatet | 2008-10-06 21:59:48 +0200 (lun. 06 oct. 2008) | 2 lignes

Move the package to the r-forge repo

------------------------------------------------------------------------
r220 | mathieu_ribatet | 2008-10-06 21:59:21 +0200 (lun. 06 oct. 2008) | 2 lignes

Move the package to the r-forge repo

------------------------------------------------------------------------
r219 | mathieu_ribatet | 2008-10-06 21:51:07 +0200 (lun. 06 oct. 2008) | 2 lignes

re-creating the pkg folder

------------------------------------------------------------------------
