2b8ab0e939
Adding error check to constructor of discrete MI calculator;
...
and making method of generating random permutations more efficient (using native library).
2013-10-22 04:08:21 +00:00
b4645584fe
Added new method generateRandomPerturbations(int, int) to RandomGenerator, and switched all computeSignificance methods to use this (since there's no need for checking for distinct perturbations really, and this is much faster). Required adding sortIndices(double[]) method to MatrixUtils
2013-10-18 00:42:12 +00:00
c7b42e38ae
Bug fix to computeSignificance() of ConditionalMutualInformationCalculator
2013-10-17 13:19:58 +00:00
8991b4176c
Minor changes to method header comments
2013-09-10 22:50:30 +00:00
7f48066b57
Predictive information calculator implementation made from active information storage
2013-08-13 04:33:18 +00:00
68b0233e28
Extended fix to handle zero-covariance conditionals for handling linear redundancies amongst variables properly in ConditionalMICalculatorMultivariateGaussian - this contributes to resolution (closed) for Issue 16
2013-07-06 18:24:13 +00:00
ce1a3b4b31
Extended fix to local cond MI values for handling linear redundancies amongst variables properly in ConditionalMICalculatorMultivariateGaussian - this resolves Issue 16
2013-07-06 13:54:02 +00:00
0907ed495e
Handle linear redundancies amongst variables properly in ConditionalMICalculatorMultivariateGaussian. Still need to extend this to the local conditional MI values
2013-07-06 11:11:56 +00:00
524b277b54
Adding minor utility functions
2013-06-18 07:31:18 +00:00
75c36dc13a
Adding common parent implementation of Conditional MI calculator from continuous to discrete variable, plus a linear-Gaussian implementation of it
2013-03-18 06:08:12 +00:00
b17ba9b05c
Minor alterations to utilities, including adding exceptions and adding to-do notes.
2013-03-18 05:55:43 +00:00
a00b24019c
Making linear-Gaussian implementation of Conditional MI calculator not normalise the data by default (after having made it a child class of the common conditional MI implementation, this had been switched on by default)
2013-03-18 05:52:36 +00:00
ddf4aaecc5
Reconditioned Conditional MI with Discrete calculator Kraskov as a child class of Conditional MI with Discrete calculator Common
2013-03-18 05:49:00 +00:00
4bd22a0984
Reconditioned Conditional MI calculator Kraskov as a child class of Conditional MI calculator Common
2013-03-11 12:55:29 +00:00
16cc8e3db2
Patched digamma function for large k; added extra debug prints to Kraskov MI calculator 2
2013-03-07 12:46:52 +00:00
cd67da53a4
Patching linear Gaussian estimators to give empirical statistical significance given a certain number of observations.
...
Changed implementation of our conditional MI calculators Kraskov via algorithm 2 to my current interpretation of how this should be done.
Added a utility for lagged covariance in MatrixUtils
2013-03-04 06:08:38 +00:00
64f6986d45
Added single array addObservations routines for Complete TE calculator discrete. Removed duplicate copyIntoMatrix method from MatrixUtils (duplicated arrayCopy)
2013-01-23 00:45:01 +00:00
afe224ab11
Added MatrixUtils.copyIntoMatrix utility function to copy part of one 2D matrix into another
2013-01-22 03:36:11 +00:00
9d36e5d9a6
Bug fix to debug prints in Kraskov Conditional MI calculators
2013-01-15 01:29:57 +00:00
6761de7e02
Added debug prints to Kraskov Conditional MI calculators
2013-01-15 01:14:13 +00:00
f466012556
Made Kraskov Conditional MI calculator implement the common Conditional MI interface
2013-01-14 12:51:34 +00:00
9b9abda733
Bug fix of computation of epsilon (box width) on conditional variable in Kraskov conditional MI calculator 2
2013-01-14 11:13:58 +00:00
3167944875
Adding interface for conditional mutual information (continuous) calculator, common methods for it, and implementation for Gaussian variables.
2013-01-12 13:43:43 +00:00
030a172d82
Patched covariance calculation, added utils, patched analytic computation of distribution of MI with linear Gaussians, altered comments
2013-01-08 05:37:48 +00:00
617880f9f5
Making Kraskov MI and higher order calculators use MAX_NORM in the marginal spaces by default (previously this had to be supplied via a property - it was the standard choice made, but was not the default). This aligns with the default specified in the Kraskov paper.
2012-12-20 10:38:03 +00:00
e789acd4dc
Adding various comments
...
Patched bug in TECalculatorMultiVariateSingleObservationsKernel whereby only k*destDimensions dimensions of destPast were being considered (this is a problem where one wants to surreptitiously condition on extra variables).
2012-12-13 14:42:40 +00:00
faad3c82b2
Patched nats -> bits error as per Issue 12.
...
Fixed comments in MI calculator header.
Added log2 static value for TECommon to make code faster.
2012-12-13 09:13:44 +00:00
29f183e91d
Altered RandomGenerator.generateDistinctRandomInts() to return int[] instead of double[].
...
Added comments in CompleteTransferEntropyCalculator to clarify the meaning of offsets from other sources.
Added comments regarding entropy calculator
2012-12-03 05:04:00 +00:00
971c0f1760
Uploaded ConditionalMutualInformationCalculator after fixing the computeSignificance function. (untested so far however)
...
Patched bad unicode character in comments for EntropyCalculatorMultivariateKozachenko
2012-10-30 06:14:44 +00:00
731c17def1
Edited comments on MatrixUtils to talk about java-octave 2D matrix conversion issues
2012-10-25 14:08:27 +00:00
0332149b03
Addressing minor compile warning
2012-10-25 01:30:47 +00:00
839da1c81f
Refactored KernelEstimatorSingleVariate to be called UniVariate
2012-10-25 01:25:10 +00:00
997fe0fbef
Added AnalyticNullDistributionComputer and refactored the reference to it in MutualInfoCalculatorMultivariateGaussian
2012-10-25 01:14:55 +00:00
6e6a21f54b
Added utility to convert a 1D matrix to a 2D container in MatrixUtils.
...
Comment changes to other classes.
2012-10-25 00:47:59 +00:00
cda8689afb
Made sure all discrete calculators use log_2 to make answers in bits (some were still using log of base).
...
Merged usual and debug methods for computing average MI in discrete calculator
Added setDebug to super InfoMeasureCalculator
Cleaned up header of InfoMeasureCalculator
Adding TODO comments for Kraskov calculators
Bug fix on MutualInfoCalculatorMultiVariateWithDiscreteKraskov - we weren't normalising incoming observations for computing local MI properly, this is fixed now.
2012-09-05 06:23:48 +00:00
d9b6393a06
Bug fix (time index on locals) for Kraskov to discrete
...
Adding interpretation of functionality to header of Gaussian to discrete (that it says how much knowing the discrete variable reduces the variance of the continuous, without taking redundancy into account)
2012-08-17 09:40:36 +00:00
07ce2d0e24
Added determinant computation via Cholesky decomposition for symmetric matrices (more efficient)
...
Utilising determinant computation via Cholesky decomposition in Gaussian entropy and MI calculators
Allowed ChannelCalculatorCommon.finaliseObservations() to throw Exceptions so that child classes can do so (in particular Gaussian MI needs to do so if it finds the variables are linearly dependent)
2012-08-17 07:28:50 +00:00
b978034f18
Added implementation for local entropy in EntropyCalculatorMultiVariateGaussian + bug fixes
...
Added MutualInfoCalculatorMultiVariateWithDiscreteGaussian
Added many javadoc comments
2012-08-16 05:40:29 +00:00
8129dbee7b
Updating javadoc comments in the code, a. to make them more explicit regarding the transfer entropy calculators, and specifically b. to show that addObservations does not append observation time series.
2012-08-08 05:33:13 +00:00
12e6f12147
Altered interal names of kernel width variables to be kernel width, instead of epsilon. This is more intuitive for users.
2012-08-07 07:49:23 +00:00
c828445784
Pulled common functionality for MutualInfoMultiVariate Kernel and Gaussian into MutualInfoMultiVariateCommon (primarily adding multiple observations and computing statistical significance).
...
MutualInfoMultiVariateKernel now computes statistical significance based on shuffling the first variable (the source) in line with the description in ChannelCalculator.
This adds ability to add multiple observations for a MutualInfoMultiVariateKernel calculator also.
2012-08-07 07:32:52 +00:00
d19045014a
Adding extra matrix utility functions to allow multivariate normal PDF to be calculated (including Cholesky decomposition, inversion of symmetric matrices), extra matrix multiplcation routines, etc.
...
Added routines in MathsUtils for uni- and multi-variate normal PDF and CDF (univariate only).
Implemented local MI for Gaussian MI. Pulled many routines into the MutualInfoMultiVariateCommon
2012-08-06 07:44:23 +00:00
6235fcbcf5
Adding analytic Chi Square measurement distribution
2012-08-01 12:12:58 +00:00
6bd3cb1661
Renamed lineargaussian package to gaussian (i got confused when i called it linear in the first place - this only is meaningful when we're working out the covariance matrix from the network structure, which isn't done here).
...
Part 2 - removing old directory, adding EntropyCalculatorMultivariate and MI calculator
2012-08-01 12:11:45 +00:00
05b3266037
Renamed lineargaussian package to gaussian (i got confused when i called it linear in the first place - this only is meaningful when we're working out the covariance matrix from the network structure, which isn't done here).
...
Part 2 - adding new files with package name change for Entropy
2012-08-01 07:43:03 +00:00
c0e98dd190
Renamed lineargaussian package to gaussian (i got confused when i called it linear in the first place - this only is meaningful when we're working out the covariance matrix from the network structure, which isn't done here). Part 1 - removing old files and creating new directory
2012-08-01 07:35:59 +00:00
9a86ee8152
Added classes for joint entropy and multivariate mutual information calculation with gaussian assumption.
...
Restructured MeasurementDistribution to have child classes EmpiricalMeasurementDistribution and AnalyticMeasurementDistribution - this resulted in many classes being altered.
Added ChiSquare distribution methods to MathsUtils.
Added more covariance methods to MatrixUtils.
2012-08-01 07:26:18 +00:00
6af2a5b0a1
Added more covariance methods to MatrixUtils.
...
Changed method signatures for initialisers for multivariate channel calculator (transfer entropy and mutual information) classes to have source dimensions before destination dimensions, in line with the addObservations (etc.) methods.
2012-07-23 07:29:52 +00:00
680724dfd3
Forgot 0.5 factor in entropy for Gaussians
2012-07-20 12:35:49 +00:00
4bb762c119
Added implementation of differential entropy for Gaussian variables.
2012-07-20 12:33:23 +00:00
joseph.lizier