ELKI COPAC implementation

Question

I tried to run COPAC ELKI implementation on the example dataset provided on the official site (mouse.csv) but I get a NullPointerException which leads me to think that there is some detail that I omit (shame on me).

The exception is the following:

Task failed java.lang.NullPointerException at de.lmu.ifi.dbs.elki.database.ids.integer.DoubleDistanceIntegerDBIDPair.compareByDistance(Unknown Source) at de.lmu.ifi.dbs.elki.distance.distanceresultlist.AbstractKNNHeap$Comp.compare(Unknown Source) at de.lmu.ifi.dbs.elki.distance.distanceresultlist.AbstractKNNHeap$Comp.compare(Unknown Source) at de.lmu.ifi.dbs.elki.utilities.datastructures.heap.TopBoundedHeap.add(Unknown Source) at de.lmu.ifi.dbs.elki.database.query.knn.LinearScanRawDoubleDistanceKNNQuery.getKNNForObjectBenchmarked(Unknown Source) at de.lmu.ifi.dbs.elki.database.query.knn.LinearScanRawDoubleDistanceKNNQuery.getKNNForObject(Unknown Source) at de.lmu.ifi.dbs.elki.database.query.knn.LinearScanRawDoubleDistanceKNNQuery.getKNNForDBID(Unknown Source) at de.lmu.ifi.dbs.elki.index.preprocessed.localpca.KNNQueryFilteredPCAIndex.objectsForPCA(Unknown Source) at de.lmu.ifi.dbs.elki.index.preprocessed.localpca.KNNQueryFilteredPCAIndex.objectsForPCA(Unknown Source) at de.lmu.ifi.dbs.elki.index.preprocessed.localpca.AbstractFilteredPCAIndex.preprocess(Unknown Source) at de.lmu.ifi.dbs.elki.index.preprocessed.localpca.AbstractFilteredPCAIndex.getLocalProjection(Unknown Source) at de.lmu.ifi.dbs.elki.index.preprocessed.localpca.AbstractFilteredPCAIndex.getLocalProjection(Unknown Source) at de.lmu.ifi.dbs.elki.algorithm.clustering.correlation.COPAC.run(Unknown Source) at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) at java.lang.reflect.Method.invoke(Method.java:616) at de.lmu.ifi.dbs.elki.algorithm.AbstractAlgorithm.run(Unknown Source) at de.lmu.ifi.dbs.elki.workflow.AlgorithmStep.runAlgorithms(Unknown Source) at [...]

This is the setting of the parameters that I use:

-dbc.in [PATH OMITTED]/mouse.csv

-algorithm clustering.correlation.COPAC

-copac.preprocessor localpca.KNNQueryFilteredPCAIndex

-copac.partitionAlgorithm gdbscan.GeneralizedDBSCAN

-dbscan.epsilon 0.00400000

-dbscan.minpts 3

Can anyone give me some info?

Answer 1

You must specify the k parameter for kNN Local PCA:

-localpca.k 10

there is a slight bug in the last ELKI release, where the parameter is incorrectly marked as optional. It is not optional. Note that k should be significantly larger than the data dimensionality to give stable PCA results.

Note that using COPAC on a low dimensional data set such as the mouse example does not make sense.