I have changed an .arff file to a .csv file in a tool in Weka. But now I can't use the arffparser as parser in ELKI.
What parser should I then use? The default is NumberVectorLabelParser. But it gives me a ArrayIndexOutOfBoundsException:
Running: -verbose -verbose -dbc.in /home/db/lisbet/Datasets/without ids/try 2/calling them .txt using another parser/Lymphography_withoutdupl_norm_1ofn.csv -dbc.parser NumberVectorLabelParser -algorithm outlier.lof.LOF -lof.k 2 -evaluator outlier.OutlierROCCurve -rocauc.positive yes
Task failed
java.lang.ArrayIndexOutOfBoundsException: 47
at de.lmu.ifi.dbs.elki.datasource.parser.NumberVectorLabelParser.getTypeInformation(NumberVectorLabelParser.java:337)
at de.lmu.ifi.dbs.elki.datasource.parser.NumberVectorLabelParser.buildMeta(NumberVectorLabelParser.java:242)
at de.lmu.ifi.dbs.elki.datasource.parser.NumberVectorLabelParser.nextEvent(NumberVectorLabelParser.java:211)
at de.lmu.ifi.dbs.elki.datasource.bundle.MultipleObjectsBundle.fromStream(MultipleObjectsBundle.java:242)
at de.lmu.ifi.dbs.elki.datasource.parser.AbstractStreamingParser.asMultipleObjectsBundle(AbstractStreamingParser.java:89)
at de.lmu.ifi.dbs.elki.datasource.InputStreamDatabaseConnection.loadData(InputStreamDatabaseConnection.java:91)
at de.lmu.ifi.dbs.elki.database.StaticArrayDatabase.initialize(StaticArrayDatabase.java:119)
at de.lmu.ifi.dbs.elki.workflow.InputStep.getDatabase(InputStep.java:62)
at de.lmu.ifi.dbs.elki.KDDTask.run(KDDTask.java:108)
at de.lmu.ifi.dbs.elki.application.KDDCLIApplication.run(KDDCLIApplication.java:60)
at [...]
My .csv file looks like this:
'Lymphatics = deformed','Lymphatics = displaced','Lymphatics = arched','Lymphatics = normal','Block_of_affere = yes','Block_of_affere = no','Bl_of_lymph_c = no','Bl_of_lymph_c = yes','Bl_of_lymph_s = no','Bl_of_lymph_s = yes','By_pass = no','By_pass = yes','Extravasates = yes','Extravasates = no','Regeneration_of = no','Regeneration_of = yes','Early_uptake_in = yes','Early_uptake_in = no','Changes_in_lym = oval','Changes_in_lym = round','Changes_in_lym = bean','Defect_in_node = lacunar','Defect_in_node = lac_central','Defect_in_node = lac_margin','Defect_in_node = no','Changes_in_node = lac_central','Changes_in_node = lacunar','Changes_in_node = no','Changes_in_node = lac_margin','Changes_in_stru = faint','Changes_in_stru = drop_like','Changes_in_stru = stripped','Changes_in_stru = coarse','Changes_in_stru = diluted','Changes_in_stru = grainy','Changes_in_stru = no','Changes_in_stru = reticular','Special_forms = vesicles','Special_forms = no','Special_forms = chalices','Dislocation_of = no','Dislocation_of = yes','Exclusion_of_no = yes','Exclusion_of_no = no',Lym_nodes_dimin,Lym_nodes_enlar,No_of_nodes_in,Outlier
1,0,0,0,1,0,1,0,1,0,1,0,1,0,1,0,1,0,1,0,0,1,0,0,0,1,0,0,0,1,0,0,0,0,0,0,0,1,0,0,1,0,1,0,0,0.333333,0.285714,no
0,1,0,0,1,0,1,0,1,0,1,0,0,1,1,0,1,0,1,0,0,1,0,0,0,1,0,0,0,1,0,0,0,0,0,0,0,0,1,0,1,0,1,0,0,0.333333,0.142857,no
There are 11 parsers available. But maybe it is my data, that is to large for the parser.
Thank you, this is a bug in the ELKI CSV parser.
It did not expect the class column to have a label.
So if your remove the ,Outlier part of your first line (or the first line completely), it should read this file just fine.
I will push a change that makes it more robust here (it will still lose the label though, because ELKI currently has support column labels for numerical columns but not for string label columns).