2714650cfa
Amending Kraskov (KSG) MI estimators to have an experimental method to provide conditional entropy of the first variable given the second. Works by removing the Kozachenko Leonenko entropy of variable 1 from the MI, using the same kNN radii as the MI estimator. Includes Unit tests to provide some initial validation.
2022-06-16 21:29:29 +10:00
07883af734
Updated printed messages with cross-platform newline character.
2021-02-09 10:23:30 +00:00
bc8c233e68
Added KSG implementation of various multivariate IT measures and unit tests.
2021-01-25 12:41:01 +00:00
a58a01fbfd
Added discrete implementation of various multivariate IT measures and unit tests.
2021-01-24 22:13:15 +00:00
f67e274819
Added Gaussian implementation of various multivariate IT measures and unit tests.
2021-01-24 22:06:39 +00:00
3222b5a629
Implementation and unit test for KSG MI calculation with new samples (both algorithms 1 and 2)
2020-10-30 16:21:10 +11:00
0606f8ab39
Cleaning up CMI Kraskov for new samples, including fixing some comments, logic of not using GPU for this, and if no conditional counting all point as a conditional match
2020-10-30 16:19:31 +11:00
d545793bc5
Patching unit test on KSG MI with discrete which tested that result changed from K=2 to 4, but the threshold for change was too large.
2019-07-25 10:14:57 +10:00
0dae162d82
Added new unit tests for TE with k=0 for Gaussian and KSG estimators
2019-07-23 11:43:55 +10:00
fba941362d
Adding further unit tests for Conditional TE (continuous-valued) to check different delays on conditional variables, and patching test of validity of calling univariate versus multivariate method signatures.
2019-07-12 14:45:23 +10:00
681c9196ee
Added methods to MatrixUtils (plus unit tests) for adding/removing/swapping sample points into existing means and covariances
2019-05-17 22:27:30 +10:00
ecec4d20e3
Added unit tests for linear dependency possibilities for issue #16
2019-03-31 23:22:55 +11:00
988d32de70
Added bug fix within recent changes to MatrixUtils Cholesky decomposition calculation, along with unit testing of the linear dependency combinations for issue #17
2019-03-31 23:21:39 +11:00
d58c393c53
Added methods to MatrixUtils to make a Cholesky decomposition of independent components only (useful for MI and conditional MI calculations), with a required update to the NonPositiveDefiniteMatrixException to include a field defining which row in the matrix was a problem. Also includes code to convert between native int[] arrays and ArrayList<Integers>, and new method signatures for column selection from matricies.
...
Also added new unit tests for the Cholesky decomposition.
2019-03-28 23:52:47 +11:00
4908276333
GPU tests and example now compute multiple surrogates.
2018-11-25 19:45:48 +02:00
8909c23fcc
Incorporated auto-embedding in the TE multivariate via conditional MI common class. Much of the code is moved up from the TE multivariate KSG estimator; now the functionality is provided for multivariate Gaussian as well. Also necessitated adding the max_cor_num_surrogates property to TE multivariate Gaussian, and took the opportunity to make much of the whitespace here consistent, and unit tests added for Gaussian (tweaked slightly for KSG)
2018-10-18 00:58:17 +11:00
a34bce25ba
Enabled auto-embedding for TE Gaussian. Did so by bringing up auto-embedding for TE functionality up from the KSG estimator to the via Conditional MI class. Also added new property to Gaussian TE for number of surrogates to use for bias correction in embedding (if not using analytic bias correction), like Gaussian AIS. This also necessitated adding an empty preFinaliseAddObservations() method to the TE Gaussian multivariate estimator, to ensure it did not yet try to autoembed until that is enabled in the multivariate calculator. Unit tests included here as well (included using more samples for TE KSG autoembedding to ensure we get the correct answer more reliably -- should probably add a stronger effect for that too).
2018-10-17 13:49:09 +11:00
efb8243ed7
Altering TE Kraskov to do auto-embedding properly when we have added multiple observations with validity vectors. Also taking the opportunity to re-engineer the internals to lay groundwork for auto-embedding function to be shared with the Gaussian estimator at a later stage. Unit tests added here as well for the above.
2018-10-15 20:56:50 +11:00
7f9b1c0812
Patching bias correction for Guassian MI and CMI (some for MI was in line with previous work for CMI, which required extending here). Unit tests included.
2018-08-27 13:17:58 +10:00
fdec72aeff
Added tracking of bias correction to ChiSquaredMeasurementDistribution, so the distribution is adjusted with bias correction. Should have been sent with previous commit for patch to CMI Gaussian bias correction
2018-08-27 00:29:59 +10:00
7fafabe451
Extending MI discrete calculator to allow different bases for each variable. This necessitates immediate removal of the (int,int) constructor (where the 2nd argument was the time difference), this will now be confusing between (base, timeDiff) and (base1, base2). In time we may bring it back, after we can be reasonably comfortable people have switched away from using (base, timeDiff).
2018-06-20 10:55:21 +10:00
5bde6f295b
Fixes issue #70 by providing setObservations(double[], int[], double[]) methods for the mixed CMI calculators (i.e. allowing univariate arrays to be passed in if dimension 1 was defined for both the continuous data and conditional). Unit test validating is included.
2018-06-19 23:08:13 +10:00
fa9db58945
In AIS Gaussian unit test, switching property names for auto embedding to now reference the super class (AIS via Mutual Info)
2018-05-16 22:44:32 +10:00
4ff4e4bfc9
Added unit tests for AIS Multivariate Gaussian calculator, including on auto-embedding, and that it functions ok for univariate data.
2018-05-16 22:43:46 +10:00
1c591925bd
Added auto-embedding to Gaussian AIS calculator, with both Ragwitz and maximum bias-corrected AIS methods. Bias-correction implemented via the Gaussian analytic surrogates. Involves defining new properties, setProperty, getProperty and preFinaliseAddObservations() to do the embedding. Unit test included also, mirroring unit testing for KSG AIS (which used linear Gaussian variables).
2018-05-15 17:35:06 +10:00
c0b9d736d7
Added methods to retrieve mean and standard deviation of the Chi^2 null distribution, as well as (empirical) unit test for this.
2018-05-13 22:02:46 +10:00
d1bba0e4c1
Adding computeLocalOfPreviousObservations() to continuous Entropy estimator interface and all underlying implementations. Added unit tests that locals should average back ok. Also fixing multivariate Entropy estimator interface to implement the univariate interface also (and fixing underlying implementations to comply)
2018-04-27 11:20:25 +10:00
c15c643a09
Added tests for new embeddings in univariate KSG TE calculator.
2018-04-22 22:25:40 +01:00
03d5e2fb2f
Added tests for new TE autoembedding methods.
2018-02-27 20:18:28 +00:00
d4c9b28973
Fixed error in KSG TE multivariate tester.
...
Tester was wrongly calling the univariate version of the algorithm using
the first column of the input.
2018-02-27 20:16:44 +00:00
438a88c97c
Added unittests for multivariate KSG TE autoembedding.
2018-02-27 17:56:42 +00:00
08f25beee6
Add tests for new overloadings in KSG mixed calc.
2018-01-12 22:23:12 +00:00
92f1154e3d
Added smoke junit tests for GPU CMI reordering.
2017-12-19 20:43:25 +00:00
4949a1079d
Added junit KSG GPU CMI tests.
2017-12-18 17:51:03 +00:00
1fb4015df5
Remove debug mode from junit GPU MI tests.
2017-12-18 17:43:15 +00:00
142fbb4541
Merge branch 'master' of github.com:jlizier/jidt
...
Merging in Pedro's mixed Kraskov changes, plus further work on GPU for Kraskov MI
2017-11-24 15:30:20 +11:00
1bb67f0840
Adding neighbour search utilities in preparation for attempting fast surrogate generation (performing multiple neighbour searches at once where we can); in particular adding a countPointsWithinR with a remapping of the point's indices. Unit test added as well for this.
2017-11-22 20:48:30 +11:00
446ac4ae59
Merge branch 'master' into gpu
2017-11-21 14:58:15 +00:00
ecda70f8e9
Hard-setting dynCorrExclTime to 0 in KSG mixed calc.
...
Previous implementation was incomplete due to a mismatch between the
sample indices in the full and marginal KdTrees. Some code has been left
to guide future development.
2017-11-20 14:47:00 +00:00
83e77e8b3a
Merge Joe's master into Pedro's master with KSG mixed changes.
2017-08-30 17:36:32 +01:00
9b7d7c70c8
Add unittests for noise level and Theiler window in KSG mixed calc.
2017-08-30 13:31:13 +01:00
9aea9c4a02
Added some tests for KSG mixed setProperty.
2017-08-30 09:14:56 +01:00
641d17ec24
Fix bug with normalisation in KSG mixed and change unittest accordingly.
2017-08-16 19:18:51 +01:00
faa98a8d79
Added more references and tests for KSG mixed calc.
2017-08-16 18:21:24 +01:00
1065662739
Added more tests to KSG mixed locals.
2017-08-16 14:51:33 +01:00
8d1d1ab0eb
Moarr tests for KSG mixed.
2017-08-16 10:24:00 +01:00
9b42911c60
Add KSG mixed unit test comparing against analytical value.
2017-08-16 01:01:31 +01:00
dcd6314082
Add a few unit tests for KSG mixed calculator.
2017-08-16 00:26:53 +01:00
ea40f48bbc
Adding new methods computePValueForGivenEstimate() and computeEstimateForGivenPValue to AnalyticMeasurementDistribution and implementing in ChiSquareMeasurementDistribution. Involves importing (and refactoring) a significantly larger chunk of commons.maths3 classes (and updating existing ones to latest 3.6.1 version for consistency). Also addresses issue #23 in changing the actual value of the ChiSquareMeasurementDistribution to be that of the information theoretic measurement rather than 2*N times it (which was the value that is actually chi squared distributed), which also necessitates changes to the Discrete and Gaussian calculators computeSignificance() methods.
2017-06-14 11:09:01 +10:00
c0a4a283fa
Added junit test for GPU MI calculation with non-zero exclusion window.
2017-05-31 04:20:38 +10:00
jlizier