Skip to content

Commit

Permalink
ANNZ 2.0.5
Browse files Browse the repository at this point in the history
  • Loading branch information
IftachSadeh committed Jun 17, 2015
1 parent f16cb3c commit 2845404
Show file tree
Hide file tree
Showing 2 changed files with 5 additions and 1 deletion.
4 changes: 4 additions & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,8 +2,12 @@

## Master version

## ANNZ 2.0.5 (17/6/2015)

- Fixed bug in `CatFormat::addWgtKNNtoTree()`, where the weight expression for the KNN trees did not include the `ANNZ_KNN_w` weight in cases of `glob.annz["addInTrainFlag"] = True`.

- Modified the condition on the MLM error estimator which is propagated to PDFs in randomized regression. Previously if an error was undefined, indicating a problem, the code would stop. Now the MLM is ignored and the code continues. The change is needed as sometimes there is a valid reason for the error to be undefined.

- Other minor modifications.

## ANNZ 2.0.4 (19/3/2015)
Expand Down
2 changes: 1 addition & 1 deletion README.md
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
# ANNZ 2.0.4
# ANNZ 2.0.5

## Introduction
ANNZ uses both regression and classification techniques for estimation of single-value photo-z (or any regression problem) solutions and PDFs. In addition it is suitable for classification problems, such as star/galaxy classification.
Expand Down

0 comments on commit 2845404

Please sign in to comment.