package/gonum
1
0
Fork 0
mirror of https://github.com/gonum/gonum.git synced 2025-10-16 20:20:41 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity
298 Commits 45 Branches 18 Tags
a377c62d9020ccadb2137782013f1e068151cb15
Commit Graph

72 Commits

Author SHA1 Message Date
kortschak
a377c62d90 stat: add simple linear regression 2016-08-05 14:59:01 +09:30
Armadilloa16
926be36d48 ROC function 
Produces a ROC curve either for all possible
cutoffs, or for n equally spaced cutoffs.
 2016-04-26 16:05:46 +09:30
kortschak
e65a547b1f {dist,distmv,sample,stat}: add missing copyright headers and package docs 2016-03-08 10:09:59 +10:30
Vladimir Chalupecky
5ff467937f Remove redundant slice types from function signatures 2015-02-10 21:32:08 +09:00
btracey
3093608919 Zeroed count in Histogram so the correct answer is returned when count != nil 2015-01-10 12:42:45 -08:00
btracey
a723bc27d8 Fix Histogram implementation. 
The former behavior of Histogram did not agree with the documentation. The documentation matched
the spirit of floats.Within, so keep the documentation and fix the behavior. This change updates
the function behavior, as well as corrects the test and the example.
 2015-01-10 10:41:15 -08:00
Jonathan J Lawlor
37fb643151 Merge pull request #19 from gonum/covariance-matrix 
CovarianceMatrix function
 2014-12-26 08:58:45 -05:00
Jonathan J Lawlor
fb2fe6268d fix weighted covariance implementation 
Weighted covariance accidentally used squared weights.  Added a test
case and fixed implementation.
 2014-12-23 21:41:15 -05:00
Jonathan J Lawlor
2ce22f9385 update Correlation godoc and comments 
as discussed
 2014-11-15 00:51:58 -05:00
Jonathan J Lawlor
007b3e00f6 fix punctuation in Variance godoc 2014-11-14 17:49:36 -05:00
Jonathan J Lawlor
1be4dbe329 improve comments in correlation 2014-11-14 15:31:11 -05:00
Jonathan J Lawlor
062f132f95 improve citation comment 2014-11-14 15:30:36 -05:00
Jonathan J Lawlor
f3c5d5f73a split out mean formula into its own line 2014-11-14 15:17:40 -05:00
Jonathan J Lawlor
9bbef05726 lens -> lengths 2014-11-14 11:32:41 -05:00
Jonathan J Lawlor
1ec3bdcfd9 Fix formatting in MeanVariance godoc 2014-11-14 11:32:25 -05:00
Jonathan J Lawlor
2a1b746881 rename mean and variance variables 2014-11-14 11:15:00 -05:00
Jonathan J Lawlor
cb9229f2d7 make return from MeanStdDev clearer 2014-11-14 11:09:44 -05:00
Jonathan J Lawlor
189f14fe6e move comment in MeanVariance 2014-11-14 11:05:39 -05:00
Jonathan J Lawlor
cbcf380d18 add comments about two-pass correction 2014-11-14 11:04:48 -05:00
Jonathan J Lawlor
19a293ddf5 remove unneeded Skew comment 2014-11-14 10:57:01 -05:00
Jonathan J Lawlor
1165d7d844 remove unneeded ExKurtosis comment 2014-11-14 10:56:40 -05:00
Jonathan J Lawlor
a766f2b2b7 comment for source of covariance algo 2014-11-14 00:07:45 -05:00
Jonathan J Lawlor
530d3d8d07 comment for source of correlation algo 2014-11-14 00:07:31 -05:00
Jonathan J Lawlor
0dfa0ae36f run go fmt 
forgot about it
 2014-11-13 23:13:13 -05:00
Jonathan J Lawlor
23e2570a5f change MomentAbout help text 2014-11-13 23:11:54 -05:00
Jonathan J Lawlor
b999ca7a54 use two pass correction in covariance 2014-11-13 23:11:53 -05:00
Jonathan J Lawlor
94330409b6 use two pass correction in correlation 2014-11-13 23:11:53 -05:00
Jonathan J Lawlor
cf4e806f4c Rename MomentAt -> MomentAbout 
Change the name of MomentAt to MomentAbout, as discussed on gonum-dev.
 2014-11-07 21:20:16 -05:00
Jonathan J Lawlor
de2c9e7595 Simplify Moment Sig 
Simplify the moment function to no longer take a mean, and implement
tests.
 2014-11-07 21:17:58 -05:00
Jonathan J Lawlor
cefae5e543 Simplify ExKurtosis Sig 
Remove mean and std from the input to ExKurtosis.
 2014-11-07 18:33:15 -05:00
Jonathan J Lawlor
943fa00fdc Simplify Skew Sig 
Simplify the Skew interface to remove dependence on mean and standard
deviation calculation.
 2014-11-07 18:23:43 -05:00
Jonathan J Lawlor
d310d7322a Simplify Correlation Sig 
This used similar code as the Covariance function.  Also, the example
was incorrect (verified via MATLAB) and had a line
`meanY := Mean(x, w)`

Which demonstrates the problem of calculating the weighted mean and
standard deviation externally.
 2014-11-07 00:09:08 -05:00
Jonathan J Lawlor
c69ec6cd62 Simplify covariance sig 
Changed covariance to remove the need to supply the means.  Also
implemented the corrected two-pass method to estimate the covariance.
 2014-11-06 23:00:27 -05:00
Jonathan J Lawlor
58e93505df Add MeanStdDev function 
This function is used to calculate both the mean and standard
deviation, which should save some calculation under most circumstances.
 2014-11-06 20:51:18 -05:00
Jonathan J Lawlor
40ba293d33 simplify sig to StdDev 
Removed the mean from the standard deviation calculation.
 2014-11-06 20:45:03 -05:00
Jonathan J Lawlor
1d2d682785 implement new MeanVariance and Variance 
Simplified Variance by removing the input mean from the sig, and
implemented a MeanVariance function which calculates the corrected
two-pass corrected sample variance.  It also calculates the mean, so it
returns that as well, which should save some calculation under most
cases.
 2014-11-06 20:40:10 -05:00
Jonathan J Lawlor
e9a80a0b06 Modify check on p to catch when p is NaN 
Changed the first check to make sure that p is not NaN.  Previously, it
would pass the check and could reach the “Impossible” panic depending
on the other inputs.  Now it produces a “stat: percentile out of
bounds” panic.
 2014-11-02 23:20:22 -05:00
Jonathan J Lawlor
1047631bc3 simplify bhattacharyyaCoeff 
Removed unnecessary multiply by zero “shortcut.”
 2014-11-02 23:00:23 -05:00
Jonathan J Lawlor
29c0aab8c7 fix mistake in the doc and sig for StdScore 
For some reason StdScore used “variance” when it should have read
“standard deviation.”  None of the calculations had to be changed.
 2014-10-30 23:53:30 -04:00
Jonathan J Lawlor
59a8dd5c51 reorder special cases in Quantile 
Moved the NaN check above the Sorted check, because Sorted is false for
slices with NaN.  This made the NaN code unreachable so I believe it
was unintentional.
 2014-10-30 23:33:26 -04:00
Jonathan J Lawlor
67819425bf reorder KS special cases for NaN 
The NaN checks were evaluated after the checks for sorted inputs, and
because slices with NaNs are unsorted, the NaN checks were unreachable.
 I’ve reordered them to be in line with the spirit of the code.
 2014-10-30 22:50:55 -04:00
Jonathan J Lawlor
46c8ade262 Reword help for KolmogorovSmirnov 
Reworded the first couple sentences to KS, which seemed to be sentence
fragments.
 2014-10-30 22:37:23 -04:00
Jonathan J Lawlor
0c7ecf8e08 fix spelling mistake 2014-10-30 20:49:32 -04:00
Jonathan J Lawlor
19f7ba742a simplify implementation of cross entropy 
Changed the implementation to match the simpler flow of Entropy.  This
will also result in fewer operations.
 2014-10-19 21:59:35 -04:00
Jonathan J Lawlor
09110229ac remove redundant check for nil 
The previous conditional covers the cases where weights == nil, so
there is no need to check the converse.
 2014-10-19 21:52:54 -04:00
Jonathan J Lawlor
f510cbd899 Handle NaN inputs in CDF 
When sort.Float64sAreSorted is called, it panics on NaN inputs.

The way the code is written, there seems to be an intent that if the
input x contains a NaN, then the function should return a NaN.  I
switched the order of the conditional statements so that this became
possible.

I also modified the test cases so that cases with NaN could be
evaluated.
 2014-10-19 18:28:50 -04:00
Jonathan J Lawlor
c9b363350b improve bhattacharyyaCoeff short circuit 
Old code seemed to be a typo.  There was no difference in result though.
 2014-10-19 17:42:06 -04:00
Jonathan J Lawlor
ff1d7c0a89 fix spelling error CumululantKinds 
There was a misspelling in the doc for CDF.
 2014-10-19 17:06:44 -04:00
btracey
4cb368c5a0 Improved comments. Fixed special case documentation of the public comment, and added better documentation to the working 
of the algorithm
 2014-09-18 18:32:23 -07:00
btracey
13e9477173 Unreachable code path panic change 2014-09-18 17:56:14 -07:00