Project details for NaN toolbox

Logo NaN toolbox 2.0

by schloegl - July 28, 2009, 13:38:20 CET [ Project Homepage BibTeX Download ]

view (16 today), download ( 3 today ), 1 subscription


The NaN-toolbox provides a number of statistics functions and machine learning methods for the use with Octave and Matlab. The functions can handle data with missing values encoded as NaNs, weighting of data samples, and multi-class classification (using a one-versus-rest scheme). There is a common interface to a number of different classification methods (including FDA, LDA, Naive Bayes, QDA, RDA, sparse classifiers, interfaces to some SVMs, regression/PLS, Wiener-Hopf).

Changes to previous version:

Initial Announcement on

BibTeX Entry: Download
URL: Project Homepage
Supported Operating Systems: Agnostic
Data Formats: Matlab
Tags: Classification, Multi Class, Machine Learning, Missing Data, Statistics, Weighting
Archive: download here

Other available revisons

Version Changelog Date

Changes in v.2.5.2 - faster version of quantile if multiple quantiles are requested - removes the dependency on ZLIB and thus - fixes "pkg install nan" for Octave on Windows - a number of minor improvements

For details see the CHANGELOG at

February 10, 2012, 11:45:52

Bug fixes and smaller improvements.

For details see the CHANGELOG at

November 11, 2011, 20:57:55
  • Default built is with OpenMP enabled

  • mex binaries (*.mexw64) for 64 bit windows/matlab included

  • *.mexw32 were built mit OpenMP enabled (increase speed on multi-core machines)

  • minor issues: include Performance_test, cumsum_skipnan.

May 18, 2011, 12:06:10
  • Parallel data processing on multiple CPU cores is supported.

  • Some experimental functions for signal processing (NANFFT, NANCONV, NANFILTER, ...) included

April 20, 2011, 12:17:17

bug fix: - cdfplot new: - bland-altmann improved: - xptopen supports on-the-fly decompression of gzip-files - makefile for building mex-files

February 14, 2011, 18:00:45

bug fix: fix deletion method for libSVM and LibLinear classifier (support for missing values fixed)

September 28, 2010, 02:06:54

xptopen supports : - reading of ARFF and STATA format - reading and writing of the SAS-XPT file format

September 8, 2010, 10:24:37

xptopen.mex: [HTML_REMOVED] reads and writes SAS Transport format (*.xpt) ttest and ttest2: [HTML_REMOVED] paired and unpaired t-test for data with missing values (useful for users of Octave, and for Matlab users without statistics toolbox)

  • train_sc: bug fix when labels are single column {-1,1} format
  • test_sc: improved docu for R.output

For more details see:

August 27, 2010, 15:49:59

This is a bug fix release.

Bug fix and performance improvement for median and kth_element.

For more details see:

July 29, 2010, 22:37:49
  • support for weighted liblinear
  • row-column deletion for data with missing values improved
  • faster median computation through the use of kth_element
  • mexw32 binaries included
  • better make/built scripts
  • a several minor improvements

Some minor bug fixes. For details see:

June 26, 2010, 01:33:32

-) Feature ranking algorithm added (fss.m) -) train_sc: {-1,+1}-encoding of classlabels supported weighted liblinear and svm source code included add "Deletion"-Mode: this enables NaN-support for training algorithms that did not have support for data with SVM (liblinear, SVM, etc.) -) str2double.mex for fast decoding of delimiter files. -) bug fixes (train_sc PLA)

Some minor bug fixes. For details see:

February 28, 2010, 22:00:21

improved accuracy in COVM and SUMSKIPNAN using Kahan's formula and extended double; fixed XVAL and a few minor bugs.

Some minor bug fixes. For details see:

November 22, 2009, 23:05:53

fixed STD and VAR: fixed empty opt did not resolve to default removed REM and MOD in order to avoid annoying warnings

Some minor bug fixes. For details see:

November 12, 2009, 15:32:55

New: load_fisheriris Changed: xval: fix ambiguity between W and NG if W is empty

Some minor bug fixes. For details see:

September 24, 2009, 15:17:11

New: fss: a feature ranking algorithm is included. cat2bin: converts categorial data into binary data Improvements: train_sc supports psvm with weighted samples, PLA and Winnow algorithm

Some minor bug fixes. For details see:

September 4, 2009, 17:07:25

Initial Announcement on

July 28, 2009, 13:38:20


No one has posted any comments yet. Perhaps you'd like to be the first?

Leave a comment

You must be logged in to post comments.