Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
Commits on Mar 20, 2015
  1. 0.9.1

    authored
  2. Remove contributors list - consult github instead, update changelog

    authored
    and there are many more nowadays which is a very good thing.
  3. Merge pull request #89 from Joker-vD/master

    authored
    Fix for jenksBreaks() to correctly pull the breaks values
  4. @Joker-vD

    Fix for an existing jenks.test plus 2 new tests.

    Joker-vD authored
    * Test "assigns correct breaks" is now named "assigns correct breaks (small gaps between classes)".
    
    * This test also used to be completely wrong: previously jenks([1, 2, 4, 5, 7, 9, 10, 20], 3) used to return [1, 2, 5, 20] - which means breaking into classes [1], [2, 4], [5, 7, 9, 10, 20]. However, such breaking has total variance (which Jenks' natural breaks algorithm is supposed to minimize) equal to 136.8. The optimal breaking is actually [1, 2, 4, 5], [7, 9, 10], [20], it has total variance equal to 14.(6), and the corresponding breaks vector is [1, 7, 20, 20].
    
    * New test "assigns correct breaks (large gaps between classes)" introduced. It tests for jenks([2, 32, 33, 34, 100], 3) to return [2, 32, 100, 100] that corresponds to breaking points into classes [2], [32, 33, 34], [100]. I believe it is self-obvious that such breaking is optimal (total variance is 0.5).
    
    * New test "assigns correct breaks (breaking N points into N classes)" introduced. It teasts for jenks([9, 10, 11, 12, 13], 5) to return [9, 10, 11, 12, 13, 13] that corresponds to breaking points into classes [9], [10], [11], [12], [13]. It's optimal because it's the only way to break five 5 points into 5 classes. The previous version of jenks() used to return nonsensical [9, null, 9, 10, 11, 13].
  5. @Joker-vD
Commits on Feb 22, 2015
Commits on Feb 7, 2015
  1. Merge pull request #84 from ericfischer/erf

    authored
    Fix rounding error in cumulative_std_normal_probability lookup.
  2. @ericfischer

    Update documentation HTML

    ericfischer authored
  3. @ericfischer
  4. @ericfischer
  5. @ericfischer

    Test that error_function and cumulative_std_normal_probability match

    ericfischer authored
    And fix my documentation error about their relationship
  6. @ericfischer

    Fix rounding error in cumulative_std_normal_probability lookup.

    ericfischer authored
    And add a numerical approximation to error_function() instead
    of only having a table lookup.
Commits on Sep 9, 2014
  1. Merge pull request #78 from abenrob/master

    authored
    quantiles doc  '>=' updates
  2. @abenrob

    quantiles doc '>=' updates

    abenrob authored
Commits on Aug 20, 2014
  1. Merge pull request #77 from jseppi/rms

    authored
    add `root_mean_square()`, ref #67
  2. @jseppi

    removed unnecessary name

    jseppi authored
  3. @jseppi

    add `root_mean_square()`, ref #67

    jseppi authored
  4. Matchup non-npms

    authored
  5. 0.9.0

    authored
  6. Update docs

    authored
  7. Pre-version

    authored
  8. Merge pull request #76 from tmcw/shuffle

    authored
    Implement Fisher-Yates shuffle for random sub-selections
Commits on Aug 18, 2014
  1. Update dependencies

    authored
  2. Merge pull request #74 from tmcw/tape

    authored
    Tape
  3. add and rename tests

    authored
Commits on Aug 17, 2014
  1. add tape dep

    authored
  2. Switch to tape

    authored
  3. Start on chunks

    authored
Commits on Jun 14, 2014
Something went wrong with that request. Please try again.