Skip to content

Releases: aschleg/hypothetical

hypothetical 0.3.1 release

17 Aug 14:07
Compare
Choose a tag to compare

This release is a quick fix for McNemar's test in the contingency module. There was some misleading literature in one
of the references which led to some incorrect changes in the previous release. This should be resolved now and an
update to the tests is coming up soon as well.

hypothetical 0.3.0 release

17 Aug 12:04
Compare
Choose a tag to compare

Version 0.3.0

Version 0.3.0 is adds much more statistical tests and methods and fixes several bugs. Please see below or the Changlelog in the root repo directory for all the new changes.

New Additions:

Factor Analysis

  • Factor Analysis
    • Several algorithms for performing Factor Analysis are available, including principal components, principal
      factors, and iterated principal factors.

Nonparametric

  • Wald-Wolfowitz Two Sample Nonparametric Runs Test
    • A nonparametric test for determining if two independent samples have been drawn from the same population or
      that they differ in any respect.

Analysis of Variance

  • Tests for Homogenity of Variance
    • Bartlett's Test for Homogenity of Variances
    • Levene's Test for Homogenity of Variances
  • Analysis of Variance
    • Van Der Waerden's (normal scores) Test

Critical Values

  • Critical Value Tables and Lookup Functions
  • D critical value (used in the Kolomogorov-Smirnov Goodness-of-Fit test) table and lookup function have been added.

Updates and changes:

Contingency

  • Chi Square Test of Dependence
    • Updates to methods in ChiSquareContingency.
      • The previous measures of association method has been removed in favor
        of specific methods for each measure of association. These include:
        • Cramer's $V$
        • Phi Coefficient, $\phi$
        • Contingency Coefficient, $C$
        • Tschuprow's Coefficient, $T$ (new)
      • The test_summary attribute of an initialized ChiSquareContingency class
        now has separate key-value pairs for each computed association measure.
  • McNemar's Test
    • The test has been fixed to use the correct cells in the 2x2 contingency table.
    • Continuity correction is now applied by default when performing the test. Setting the parameter continuity=False
      will perform the non-continuity corrected version of the test.

Nonparametric

  • Median Test

    • A multiple comparisons posthoc test is now available.
  • Many updates to documentation and docstrings.