This toolbox is meanly ported from the nan-toolbox for matlab/octave.
Comparison of thrownan and nan_toolbox
======================================
Performance:
size 5000 x 10: sumskipnan: 0.001 sec, sum(thrownan()) 0.002 sec
size 5000 x 50: sumskipnan: 0.005 sec, sum(thrownan()) 0.014 sec
size 5000 x 100: sumskipnan: 0.010 sec, sum(thrownan()) 0.025 sec
size 5000 x 150: sumskipnan: 0.015 sec, sum(thrownan()) 0.039 sec
size 5000 x 200: sumskipnan: 0.018 sec, sum(thrownan()) 0.051 sec
sumskipnan is 2.8 times faster then sum(thrownan()).
The second disadvantage of thrownan is, that a matrix is transformed into a
vector:
-->size(d)
ans =
5000. 200.
-->size(thrownan(d))
ans =
1000000. 1.
Changelog:
=========
1.3.2
- nan_flix is obsolete as is the same as interp1(D,X,'linear');
- using distfun toolbox for norminv and betainv
- some statistical test are added :
- "nan_ttest"
- "nan_ttest2"
- "nan_ztest"
- "nan_welchtest"
- "nan_signtest"
- "nan_signrank"
- "nan_barttest"
- "nan_anova"
- nan_probplot added
- removed bug in nan_conv2
1.3.1
- nan_qqplot,nan_linkage,nan_pdist, nan_quareform added
- nan_boxplot improved
- bugs in nan_confusionmat, nan_kappa and nan_partest removed
- sum minor bug fixes
- some unit tests are added
- replace mtlb_repmat by repmat
1.3
- removed function (available in the atom modul libsvm):
- nan_svmrocplot.sci
- train
- predict
- svmtrain
- svmpredict
- readsparse
- writesparse
- the libsvm and the nan-toolbox are compatible now!
1.2.3
- some small bugfixes
- all sci_gateway functions can be compliled for Windows
- xtpopen works again
1.2.2
- prepare toolbox for scilab 5.4
- some bug fixes
1.2.1
- remove xtpopen, now it should build under linux
1.2
- rename every function to nan_*
1.1
- some functions are added
- help files and demos are improved
1.0
- first release of NaN - toolbox
Functions:
===========
Data Correlation and Covariance
"nan_covm"
"nan_ecovm"
"nan_decovm"
"nan_xcorr"
"nan_xcorr2"
"nan_conv"
"nan_conv2"
"nan_xcov"
"nan_xcovf"
"nan_conv2nan"
"nan_cor"
"nan_cov"
"nan_corrcoef"
"nan_corrcov"
"nan_rankcorr"
"nan_partcorrcoef"
"nan_tiedrank"
Descriptive Statistics
"nan_mean"
"nan_mean"
"nan_var"
"nan_std"
"nan_median"
"nan_iqr"
"nan_trimmean"
"nan_center"
"nan_geomean"
"nan_detrend"
"nan_moment"
"nan_range"
"nan_medAbsDev"
"nan_sem"
"nan_meansq"
"nan_sumsq"
"nan_mad"
"nan_rms"
"nan_statistic"
"nan_skewness"
"nan_kurtosis"
"nan_zscore"
"nan_zScoreMedian"
"nan_meanAbsDev"
"nan_harmmean"
"nan_spearman"
"nan_ranks"
"nan_quantile"
"nan_prctile"
"nan_percentile"
"nan_trimean"
"nan_meandev"
"nan_ecdf"
"nan_coef_of_variation"
"nan_hist2res"
"nan_histo"
"nan_histo2"
"nan_histo3"
"nan_histo4"
"nan_histc"
"nan_grpstats"
"nan_y2res"
"nan_cumsum"
"nan_filter"
"nan_pdist"
"nan_linkage"
"nan_squareform"
statistical Visualization
"nan_bland_altmann"
"nan_cdfplot"
"nan_gscatter"
"nan_boxplot"
"nan_normplot"
"nan_andrewsplot"
"nan_hist"
"nan_ecdfhist"
"nan_fscatter3"
"nan_plotmatrix"
"nan_gplotmatrix"
"nan_parallelcoords"
"nan_errorb"
"nan_errorbar"
"nan_nhist"
"nan_qqplot"
hypothesis Tests
"nan_ttest"
"nan_ttest2"
Classification
"nan_train_sc"
"nan_test_sc"
"nan_classify"
"nan_xval"
"nan_kappa"
"nan_train_lda_sparse"
"nan_fss"
"nan_cat2bin"
"nan_row_col_deletion"
"nan_mahal"
"nan_rocplot"
"nan_svmrocplot"
"nan_confusionmat"
"nan_partest"
cluster Analysis
"nan_kmeans"
utility functions
"sumskipnan"
"nan_postpad"
"nan_prepad"
"nan_fft"
"nan_ifft"
"flag_nans_occured"
"flag_impl_skip_nan"
"flag_accuracy_level"
"flag_impl_significance"
"nan_grp2idx"
"nan_ismember"
"nan_unique"
"nan_mgrp2idx"
"nan_accumarray"
"nan_flix"
"str2array"