GPAttributeGeneration filter on the command line

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

GPAttributeGeneration filter on the command line

Capital Markets
Have no problem using the GPAttributeGeneration filter within the Explorer but for some reason it fails on the regular command line (which I need to use in order to run it in batch and produce filtered learning and training sets i.e. -b, -i, -o, -r, -s command arguments).

Batch, non-batch, class = 'last', class = 1, 100 instances, 10000 instances, no approach remove the following error:


./run.sh weka.filters.supervised.attribute.gpattributegeneration.GPAttributeGeneration -c last -b -i ./learn.arff -o ./gplearn.arff -r ./test.arff -s ./gptest.arff

weka.core.WekaException: weka.filters.supervised.attribute.gpattributegeneration.GPAttributeGeneration: Not enough training instances (required: 1, provided: 0)!
     at weka.core.Capabilities.test(Capabilities.java:1401)
     at weka.core.Capabilities.test(Capabilities.java:1208)
     at weka.core.Capabilities.testWithFail(Capabilities.java:1506)
     at weka.filters.Filter.testInputFormat(Filter.java:474)
     at weka.filters.Filter.setInputFormat(Filter.java:492)
     at weka.filters.Filter.batchFilterFile(Filter.java:1328)
     at weka.filters.Filter.runFilter(Filter.java:1409)
     at weka.filters.supervised.attribute.gpattributegeneration.GPAttributeGeneration.main(GPAttributeGeneration.java:116)


./run.sh weka.filters.supervised.attribute.gpattributegeneration.GPAttributeGeneration -c last -i ./learn.arff -o ./gplearn.arff

weka.core.WekaException: weka.filters.supervised.attribute.gpattributegeneration.GPAttributeGeneration: Not enough training instances (required: 1, provided: 0)!
     at weka.core.Capabilities.test(Capabilities.java:1401)
     at weka.core.Capabilities.test(Capabilities.java:1208)
     at weka.core.Capabilities.testWithFail(Capabilities.java:1506)
     at weka.filters.Filter.testInputFormat(Filter.java:474)
     at weka.filters.Filter.setInputFormat(Filter.java:492)
     at weka.filters.Filter.filterFile(Filter.java:1124)
     at weka.filters.Filter.runFilter(Filter.java:1411)
     at weka.filters.supervised.attribute.gpattributegeneration.GPAttributeGeneration.main(GPAttributeGeneration.java:116)


Again, Explorer has no problem applying this filter.  I've ensured the command line is using *exactly* the same weka.jar and ~/wekafiles packages so it's a bit weird. Bug? or am I missing something?

Cheers for any help folks.
_______________________________________________
Wekalist mailing list
Send posts to: [hidden email]
List info and subscription status: https://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Reply | Threaded
Open this post in threaded view
|

Re: GPAttributeGeneration filter on the command line

Eibe Frank-2
Administrator
Yes, looks like there is a bug in the released version of GPAttributeGeneration. Unfortunately, the standard option to turn off capabilities checking is also missing, so capabilities checking can’t be turned off.

I’ve filed an issue on Github as we don’t maintain this package:

https://github.com/colinnuk/GPAttributeGeneration/issues/1

Cheers,
Eibe

> On 4/05/2017, at 1:07 AM, Capital Markets <[hidden email]> wrote:
>
> Have no problem using the GPAttributeGeneration filter within the Explorer but for some reason it fails on the regular command line (which I need to use in order to run it in batch and produce filtered learning and training sets i.e. -b, -i, -o, -r, -s command arguments).
>
> Batch, non-batch, class = 'last', class = 1, 100 instances, 10000 instances, no approach remove the following error:
>
>
> ./run.sh weka.filters.supervised.attribute.gpattributegeneration.GPAttributeGeneration -c last -b -i ./learn.arff -o ./gplearn.arff -r ./test.arff -s ./gptest.arff
>
> weka.core.WekaException: weka.filters.supervised.attribute.gpattributegeneration.GPAttributeGeneration: Not enough training instances (required: 1, provided: 0)!
>     at weka.core.Capabilities.test(Capabilities.java:1401)
>     at weka.core.Capabilities.test(Capabilities.java:1208)
>     at weka.core.Capabilities.testWithFail(Capabilities.java:1506)
>     at weka.filters.Filter.testInputFormat(Filter.java:474)
>     at weka.filters.Filter.setInputFormat(Filter.java:492)
>     at weka.filters.Filter.batchFilterFile(Filter.java:1328)
>     at weka.filters.Filter.runFilter(Filter.java:1409)
>     at weka.filters.supervised.attribute.gpattributegeneration.GPAttributeGeneration.main(GPAttributeGeneration.java:116)
>
>
> ./run.sh weka.filters.supervised.attribute.gpattributegeneration.GPAttributeGeneration -c last -i ./learn.arff -o ./gplearn.arff
>
> weka.core.WekaException: weka.filters.supervised.attribute.gpattributegeneration.GPAttributeGeneration: Not enough training instances (required: 1, provided: 0)!
>     at weka.core.Capabilities.test(Capabilities.java:1401)
>     at weka.core.Capabilities.test(Capabilities.java:1208)
>     at weka.core.Capabilities.testWithFail(Capabilities.java:1506)
>     at weka.filters.Filter.testInputFormat(Filter.java:474)
>     at weka.filters.Filter.setInputFormat(Filter.java:492)
>     at weka.filters.Filter.filterFile(Filter.java:1124)
>     at weka.filters.Filter.runFilter(Filter.java:1411)
>     at weka.filters.supervised.attribute.gpattributegeneration.GPAttributeGeneration.main(GPAttributeGeneration.java:116)
>
>
> Again, Explorer has no problem applying this filter.  I've ensured the command line is using *exactly* the same weka.jar and ~/wekafiles packages so it's a bit weird. Bug? or am I missing something?
>
> Cheers for any help folks.
> _______________________________________________
> Wekalist mailing list
> Send posts to: [hidden email]
> List info and subscription status: https://list.waikato.ac.nz/mailman/listinfo/wekalist
> List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

_______________________________________________
Wekalist mailing list
Send posts to: [hidden email]
List info and subscription status: https://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html