R-Forge Project News

analogue 0.6-22 uploaded to CRAN

Gavin Simpson — 2009-11-07T16:33:28+00:00

A new version of analogue was pushed to CRAN a few days ago. This release was precipitated by the release of R 2.10.0, and features the following key additions and fixes:

* WA models can now make full use of tolerance down-weighting, including cross validation, and all WA models benefit from key computations being conducted in faster C code;

* tran: now has a formula method and includes new transformations; power, 4th root and log-ratio transformations. tran() now also provides row (sample) centring;

* roc: For large problems the calculation of AUC and its standard error could overflow the largest number R currently handles. roc() now has two new arguments, 'thin' and 'max.len', which allow the number of points on the ROC curve to be thinned to a smaller number, which should allow the computations to be performed.

* Stratiplot: can now reverse the y-axis limits and has option to sort the variables plotted according to their weight average for the 'y' variable (to emphasis compositional change in 'y') or a further supplied variable.

* join: the value used to replace NAs can now be set by the user, and the inner and left joins as well as the default outer join.

See the ChangeLog for more detailed information.

Gavin

analogue 0.6-8 now on CRAN

Gavin Simpson — 2009-05-07T08:38:23+00:00

A new version of analogue is now on CRAN. This release was prompted by an request from Kurt Hornik to make the CITATION file more optimal to fit in with CRAN's automated features for presenting packages to the wider world. The new version fixes the CITATION file.

As I wasn't planning on releasing just yet, I just pushed the current devel version out, which brings with it several new features and a couple of bug fixes:

Version 0.6-8:

* residLen: new function to compute squared residual length
diagnostic for passive samples in a constrained ordination. Used as
a test of whether core samples are well fitted in a transfer function
model.

Several utility functions to compute fitted values from an ordination
and corresponding residual lengths are provided, which me be useful
for authors of other functions.

'plot' and 'hist' methods produce density plots and histograms for
'residLen' objects using base graphics.

'densityplot' and 'histogram' methods for 'residLen' objects using
Lattice graphics.

* stdError: new function to compute the weighted standard deviation of
the environment values for the k closest analogues in MAT models. This
can be used as an uncertainty measure for MAT fitted values or transfer
function predictions.

Methods are available for 'mat' and 'predict.mat'.

* predict.mat: now returns the dissimilarity matrix between the training
set samples and samples in 'newdata'.

* getK: new method for 'predict.mat'.

* CITATION: updated as per request from Kurt Hornik and CRAN.

Version 0.6-7:

* optima, tolerance: new methods to coerce objects of these classes
to data frames.

* distance: method = "kendall" was incorrectly computing the min of
the x and y components in the dissimilarity.

analogue 0.6-6 upload to CRAN

Gavin Simpson — 2009-03-10T17:59:29+00:00

A new version of analogue has been uploaded to CRAN. It contains lots of new functionality and fixes, especially to the ROC curve functions.

See the ChangeLog for details of the fixes and improvements.

For version 0.7-0 I hope to tidy the package and help up a bit, add extra documentation and vignettes, and add a proper NEWS file.

G

analogue 0.5-2 uploaded to CRAN

Gavin Simpson — 2008-05-13T06:59:46+00:00

A new version of analogue (0.5-2) has been uploaded to CRAN. This version includes fixes one serious bug:

* wa: classical deshrinking did not work, but returned the original 'env' variable.

A few additional features are also included in this release:

* wa: implemented deshrink = "none".
* wa: deshrink = "expanded" is now public and user-callable.
* join: now checks for inherits(foo, "data.frame") to confirm if all objects to join are (or inherit from) data frames. This allows join to work on objects of class "join" when 'split = FALSE' is used.

This version should be on your favourite CRAN mirror in the next few days.

Gavin

analogue 0.5-1 bug-fix release

Gavin Simpson — 2008-04-26T12:17:55+00:00

A new version of analogue was uploaded to CRAN this afternoon that fixes a number of bugs in version 0.5-0:

1) predict.wa not storing some aspects of the model fit
2) residuals were calculated back-to-front in some functions; the sign of residuals will be different now in wa(), predict.wa() and several of the mat() functions

The internal deshrinking functions were very much streamlined and improved, thanks to suggestions from Jari Oksanen. Jari has also joined the analogue development team and becomes an author of the package.

Gavin

analogue: New Version 0.5-0 submitted to CRAN

Gavin Simpson — 2008-04-14T08:00:38+00:00

Earlier today I packaged up a new version of analogue and submitted it to CRAN.

The major changes in this release are:

1) Support for weighted averaging transfer function models, and

2) Inclusion of a vignette describing how to use analogue for MAT transfer functions and analogue matching.

There were a few minor bug fixes and tweaks.

The package should make its way to CRAN mirrors over the next few days. Feedback on the new functions would be appreciated.

analogue 0.4-3 uploaded to CRAN

Gavin Simpson — 2007-10-14T14:54:34+00:00

Version 0.4-3 of analogue has been uploaded to CRAN. This version improves handling of factors in species matrices in join(). distance() is now more robust and handles factors more strictly than previous versions.