Summary of important user-visible changes for statistics 1.8.2

Important Notice:

1. Minimum datatypes package version required is now 1.2.0
2. Incompatibility with the nan and tablicious packages

News:

• Improved support for new data types
• Bug fixes in few functions
• Improved documentation

Summary of important user-visible changes for statistics 1.8.1

Important Notice:

1. Minimum datatypes package version required is now 1.1.8
2. Incompatibility with the nan and tablicious packages

News:

• New dummyvar and parseWilkinsonFormula functions
• Improved support for new data types in several functions
• Several bug fixes related to various issues

Summary of important user-visible changes for statistics 1.8.0

Important Notice:

1. Minimum Octave version required is now Octave 9.1.0
2. Additional dependency to the datatypes package
3. Incompatibility with the nan and tablicious packages

News:

• New makima function
• Updated libsvm library to version 3.36
• Updated documentation and many new demos, especially for class methods and properties

Summary of important user-visible changes for statistics 1.7.7

Important Notice:
mean, median, std, var, and mad shadow core Octave's functions before Octave 9.1
incompatibility with the nan package

News:

• Updated documentation #263, #285
• Minor fixes in cvpartition and pca #268, #269, #279

Summary of important user-visible changes for statistics 1.7.6

Important Notice:
mean, median, std, var, and mad shadow core Octave's functions before Octave 9.1
incompatibility with the nan package

News:

• There are 2 new functions (factoran and multiway) introduced in this release.
• Updated documentation and demos for distribution objects.
• cvpartition and crossval are now fully MATLAB compatible.
• Several bug fixes related to various issues #232, #226
• ClassificationKNN can use either 'exhaustive' or 'kdtree' search methods for prediction.

Summary of important user-visible changes for statistics 1.7.5

Important Notice:
mean, median, std, var, and mad shadow core Octave's functions before Octave 9.1
incompatibility with the nan package

News:

• There are 2 new functions and 4 new classdefs introduced in this release.
• Several bug fixes related to various issues #183, #199, #200, #203, #204, #212, #215
• Fixed the slow kd-tree implementation in knnsearch and rangesearch functions.

Note!

1. The old style cvpartition class has been deprecated in favor of the newly introduced classdef implementation. There is a limited brake in backwards compatibility, so old code might need some tweaking before working again. The new cvpartition has many new features and is mostly MATLAB compatible.
2. crossval needs attention, see issue #217.

Summary of important user-visible changes for statistics 1.7.4

Important Notice:
mean, median, std, var, and mad shadow core Octave's functions before Octave 9.1
incompatibility with the nan package

News:

• Minimum Octave version required is 8.1.
• Updates in ClassificationDiscriminant, ClassificationKNN, and ClassificationNeuralNetwork classes.
• Minor fixes related to issues #174, #178, #179.

Summary of important user-visible changes for statistics 1.7.3

Important Notice:
mean, median, std, var, and mad shadow core Octave's functions before Octave 9.1
incompatibility with the nan package

News:
New glmval function. Updated functionality in glmfit function.

Summary of important user-visible changes for statistics 1.7.2

Important Notice:
mean, median, std, var, and mad shadow core Octave's functions before Octave 9.1
incompatibility with the nan package

News:
Minor fixes related to issues #168, #171, #172. Added parallel support for MacOS in fcnnpredict and fcnntrain functions.

Summary of important user-visible changes for statistics 1.7.1

Important Notice:
mean, median, std, var, and mad shadow core Octave's functions before Octave 9.1
incompatibility with the nan package

News:
Minor fixes related to Classification and Regression classdefs (see issues #160, #169). Added support for probability distribution objects in kstest function (see issues #164, #165, #167).