natas_hw
/
jidt
mirror of https://github.com/jlizier/jidt
0
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Fixing handling of data normalisation in the KNN computation methods

This commit is contained in:
Joseph Lizier 2025-04-01 23:09:09 +11:00
parent 333a00fc07
commit d773508cf9
2 changed files with 35 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
java/source/infodynamics/measures/continuous/kraskov/ConditionalMutualInfoCalculatorMultiVariateKraskov.java
View File

@ -960,6 +960,22 @@ public abstract class ConditionalMutualInfoCalculatorMultiVariateKraskov
public double[] kNNDistancesForNewSamples(int startTimePoint, int numTimePoints,
double[][] newStates1, double[][] newStates2, double[][] newCondStates) throws Exception {
// Do normalisation of the incoming data if required:
double[][] states1ToUse, states2ToUse, condStatesToUse;
if (normalise) {
states1ToUse = MatrixUtils.normaliseIntoNewArray(newStates1, var1Means, var1Stds);
states2ToUse = MatrixUtils.normaliseIntoNewArray(newStates2, var2Means, var2Stds);
if (dimensionsCond != 0) {
condStatesToUse = MatrixUtils.normaliseIntoNewArray(newCondStates, condMeans, condStds);
} else {
condStatesToUse = null;
}
} else {
states1ToUse = newStates1;
states2ToUse = newStates2;
condStatesToUse = newCondStates;
}
double[] kNNdistances = new double[numTimePoints];
for (int t = startTimePoint; t < startTimePoint + numTimePoints; t++) {
@ -967,7 +983,7 @@ public abstract class ConditionalMutualInfoCalculatorMultiVariateKraskov
// finding the kth closest neighbour for the new sample:
PriorityQueue<NeighbourNodeData> nnPQ =
kdTreeJoint.findKNearestNeighbours(k,
new double[][] {newStates1[t], newStates2[t], newCondStates[t]});
new double[][] {states1ToUse[t], states2ToUse[t], condStatesToUse[t]});
// First element in the PQ is the kth NN,
// and epsilon = kthNnData.distance
NeighbourNodeData kthNnData = nnPQ.poll();
@ -997,7 +1013,7 @@ public abstract class ConditionalMutualInfoCalculatorMultiVariateKraskov
"the univariate kNNDistancesForNewSamples(int, int, double[],double[],double[][]) " +
"method is called");
}
return computeLocalUsingPreviousObservations(
return kNNDistancesForNewSamples(startTimePoint, numTimePoints,
MatrixUtils.reshape(newStates1, newStates1.length, 1),
MatrixUtils.reshape(newStates2, newStates2.length, 1),
newCondStates);
@ -1023,7 +1039,7 @@ public abstract class ConditionalMutualInfoCalculatorMultiVariateKraskov
"the univariate kNNDistancesForNewSamples(int, int, double[],double[],double[]) " +
"method is called");
}
return computeLocalUsingPreviousObservations(
return kNNDistancesForNewSamples(startTimePoint, numTimePoints,
MatrixUtils.reshape(newStates1, newStates1.length, 1),
MatrixUtils.reshape(newStates2, newStates2.length, 1),
MatrixUtils.reshape(newCondStates, newCondStates.length, 1));

18
java/source/infodynamics/measures/continuous/kraskov/MutualInfoCalculatorMultiVariateKraskov.java
View File

@ -1158,6 +1158,20 @@ public abstract class MutualInfoCalculatorMultiVariateKraskov
public double[] kNNDistancesForNewSamples(int startTimePoint, int numTimePoints,
double[][] newVar1Observations, double[][] newVar2Observations) throws Exception {
double[][] states1ToUse, states2ToUse;
if (normalise) {
states1ToUse = MatrixUtils.normaliseIntoNewArray(newVar1Observations, sourceMeansBeforeNorm, sourceStdsBeforeNorm, 0, newVar1Observations.length-timeDiff);
states2ToUse = MatrixUtils.normaliseIntoNewArray(newVar2Observations, destMeansBeforeNorm, destStdsBeforeNorm, timeDiff, newVar2Observations.length-timeDiff);
} else {
if (timeDiff > 0) {
states1ToUse = MatrixUtils.selectRows(newVar1Observations, 0, newVar1Observations.length-timeDiff);
states2ToUse = MatrixUtils.selectRows(newVar2Observations, 0, newVar2Observations.length-timeDiff);
} else {
states1ToUse = newVar1Observations;
states2ToUse = newVar2Observations;
}
}
double[] kNNdistances = new double[numTimePoints];
for (int t = startTimePoint; t < startTimePoint + numTimePoints; t++) {
@ -1165,7 +1179,7 @@ public abstract class MutualInfoCalculatorMultiVariateKraskov
// finding the kth closest neighbour for the new sample:
PriorityQueue<NeighbourNodeData> nnPQ =
kdTreeJoint.findKNearestNeighbours(k,
new double[][] {newVar1Observations[t], newVar2Observations[t]});
new double[][] {states1ToUse[t], states2ToUse[t]});
// First element in the PQ is the kth NN,
// and epsilon = kthNnData.distance
NeighbourNodeData kthNnData = nnPQ.poll();
@ -1195,7 +1209,7 @@ public abstract class MutualInfoCalculatorMultiVariateKraskov
"the univariate kNNDistancesForNewSamples(int, int, double[],double[]) " +
"method is called");
}
return computeLocalUsingPreviousObservations(
return kNNDistancesForNewSamples(startTimePoint, numTimePoints,
MatrixUtils.reshape(newVar1Observations, newVar1Observations.length, 1),
MatrixUtils.reshape(newVar2Observations, newVar2Observations.length, 1));
}