Commit Graph

197 Commits

Author SHA1 Message Date
Joseph Lizier ec8e7f848f CMI common estimators utilising underlying nearest neighbour searchers holding different data sets separately (for dynamic correlation exclusion) 2023-10-09 13:15:56 +11:00
jlizier 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
Joseph Lizier 0e33393cdc
Merge pull request #86 from pmediano/master
Thanks Pedro, and sorry to take so long to attend to this
2021-11-15 12:08:57 +11:00
Joseph Lizier 21ae0bdeaa Fixing logic error in printing whether GPU calculation was falling back to CPU 2021-10-08 11:28:37 +11:00
Pedro Mediano 07883af734 Updated printed messages with cross-platform newline character. 2021-02-09 10:23:30 +00:00
Pedro Mediano bc8c233e68 Added KSG implementation of various multivariate IT measures and unit tests. 2021-01-25 12:41:01 +00:00
Joseph Lizier 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
Joseph Lizier 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
Joseph Lizier fce1b34a5c Adding NORMALISE and NOISE_LEVEL_TO_ADD properties to all conditional mutual information calculators for continuous-valued data (extends this capability from only KSG to Gaussian as well) 2019-04-03 00:12:00 +11:00
Joseph Lizier b1437452e4 Adding NORMALISE and NOISE_LEVEL_TO_ADD properties to all mutual information calculators for continuous-valued data (extends this capability from only KSG to Gaussian and kernel as well, the latter already had NORMALISE) 2019-04-02 12:38:17 +11:00
Pedro Martinez Mediano c616ba55d4 Added link to wiki in GPU loader error message. 2018-11-25 19:25:33 +02:00
jlizier 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
jlizier da418c31a0 Added new auto-embedding capability for TE Gaussian into the AutoAnalyser GUI demo. Also added new auto-embedding drop down option of Max corrected AIS + TE for KSG as well, and debugged the getProperty returning of the number of nearest neighbours for Ragwitz criteria now that this occurs in the common superclass instead of in KSG (and gave this a reasonable default of 4) 2018-10-17 14:23:20 +11:00
jlizier 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
jlizier 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
jlizier 2ff6d8aeec Added getProperty for AIS Kraskov multivariate. Need to re-engineer this class for proper code re-use anyway. 2018-10-15 14:38:46 +11:00
jlizier 908cf675a8 Added new constructor for TE Kraskov to pass in KSG algorithm number as an integer (It's not likely that it will be needed, but is there for completeness to match the AIS one) 2018-08-24 14:48:11 +10:00
jlizier 373ebbdd56 Enabling AIS KSG calculator to have underlying KSG algorithm changed via the "ALG_NUM" property 2018-08-24 14:47:21 +10:00
jlizier 643212840a Making indenting style consistent through file 2018-05-18 08:56:47 +10:00
jlizier e8603b08f8 Moving auto-embedding for AIS (Ragwitz and maximum bias-corrected AIS methods) from out of Gaussian and KSG calculators and into the common AIS calculator via MI class. The common implementation then allows us to introduce auto-embedding for the AIS kernel estimator as well (included requiring adding a property for the number of surrogates to use to compute the bias for this one).
These changes now incorporate auto-embedding for all AIS continuous calculators, partially addressing issue #38
2018-05-16 14:20:08 +10:00
jlizier 9552091fca Adding Javadoc on header for neighbour counts method in KSG MI estimator 1 2018-05-15 17:28:39 +10:00
jlizier 860d43dbca Adding additional error message for Exception during KSG Active Info Storage auto-embedding 2018-05-07 23:15:33 +10:00
jlizier bf0062b902 Switching new debug method in KSG conditional MI calculator 1 to be publicly accessible 2018-04-26 11:43:04 +10:00
jlizier 7d76f82bb2 Adding partialNeighbourCountFromObservations() method to KSG conditional MI calculator 1 for debugging investigations into the neighbour counts 2018-04-26 11:24:56 +10:00
Pedro Martinez Mediano df8eec27cd Added MAX_CORR_AIS_AND_TE embedding to univariate KSG TE calculator. 2018-04-22 22:25:05 +01:00
Pedro Martinez Mediano f8bca3275d Fixed docs for new MAX_CORR_AIS_AND_TE embedding method. 2018-04-22 20:45:06 +01:00
Pedro Martinez Mediano 02dad26a94 Added brief mention of AUTO_EMBED_METHOD_MAX_CORR_AIS_AND_TE to javadocs. 2018-02-27 20:29:39 +00:00
Pedro Martinez Mediano 5f6671aa2b Added AUTO_EMBED_METHOD_MAX_CORR_AIS_AND_TE to multivariate KSG TE. 2018-02-27 20:18:04 +00:00
Pedro Martinez Mediano 8be6a037af Added auto-embedding methods to multivariate KSG TE calculator. 2018-02-25 19:23:09 +00:00
Pedro Martinez Mediano ce4b1599c6 Added MAX_CORR_AIS embedding method to multivariate KSG AIS. 2018-02-25 18:57:15 +00:00
Pedro Martinez Mediano 64f1142b40 Added argument to choose which variable to reorder in GPU CMI code. 2017-12-19 20:42:58 +00:00
Pedro Martinez Mediano 65ac0e9ff8 Removed unnecesary overloaded method in KSG CMI calc. 2017-12-19 20:41:11 +00:00
Pedro Martinez Mediano 2f1cd192b8 Fixed small typo in Java KSG CMI calculator. 2017-12-18 17:50:37 +00:00
Pedro Martinez Mediano d0304ae0c4 Added Java-side JNI functions and ensureKdTreesConstructed to KSG CMI. 2017-12-12 11:19:40 +00:00
Pedro Martinez Mediano f4ca442de4 Now KSG MI calc does not build KdTrees if GPU is going to be used. 2017-12-12 11:10:50 +00:00
Pedro Martinez Mediano d010f32635 Updated getProperty and removed spurious dependency in KSG MI calc. 2017-12-12 11:04:32 +00:00
jlizier 61f4bf3276 Adding getProperty lines to handle USE_GPU property for KSG MI calculator 2017-11-27 09:26:19 +11:00
Pedro Martinez Mediano 446ac4ae59 Merge branch 'master' into gpu 2017-11-21 14:58:15 +00:00
Pedro Martinez Mediano ed71502bc2 Remove old cudaLibraryLoad code and unnecessary dependencies. 2017-08-31 16:22:12 +01:00
jlizier b2e3ae1ca0 Added missing getProperty() method to Conditional TE calculator Kraskov 2017-08-23 00:19:09 +10:00
jlizier 086bee5945 Making continuous MultiInfoCalculator classes implement the InfoMeasureCalculatorContinuous interface. For the interfaces, this means removing methods where duplicated. For implementing classes, this means adding the missing methods. 2017-08-18 23:42:43 +10:00
jlizier a2b9939892 Merge branch 'master' of github.com:jlizier/jidt
to combine local changes with Pedro's GPU pull
2017-06-13 12:08:54 +10:00
jlizier 77f4c1afd5 Added partialNeighbourCountFromObservations() method for Leo's analytic work on neighbour counts 2017-06-13 12:04:41 +10:00
Pedro Martinez Mediano d0b35b3db8 Added NativeUtils for native lib loading and included libKraskov.so in jar. 2017-06-01 20:34:23 +10:00
Pedro Martinez Mediano 834a3ec2c2 Relocated GPU code and modified build and load paths accordingly. 2017-06-01 17:22:42 +10:00
Pedro Martinez Mediano 2a203faa23 Added support for Theiler exclusion window in JNI code. 2017-05-31 04:19:49 +10:00
Pedro Martinez Mediano bc44e4c2fb Main workflow now operates using only device functions.
This saves time and energy by keeping all data in te GPU, rather than
transfering it back and forth all the time.
2017-05-31 02:38:44 +10:00
Pedro Martinez Mediano 85dd28d13b Separated digamma and reduction into host and device functions. 2017-05-31 02:36:45 +10:00
Pedro Martinez Mediano 7a89dd7908 Modified handling of normFunction and replaced signallength with signalLength. 2017-05-31 02:35:11 +10:00
Pedro Martinez Mediano 6c25ea93ef Added functions to allocate and free all the GPU memory at once. 2017-05-31 02:31:37 +10:00