Re: MISMO Does not implement WeightedInstancesHandler Interface

> I am attempting to build a MISMO classifier from an .arff formatted dataset of multiple-instance data which do not include any instance weights. I ultimately want to access this classifier through the python-weka-wrapper3 python module, but have been encountering a Java Exception stating MISMO does not implement the WeightedInstancesHandler interface. Aside from being inconsistent with the documentation (https://weka.sourceforge.io/doc.stable/weka/classifiers/mi/MISMO.html) and what I saw when actually looking at the MISMO.java class, I would not have expected any issues related to weights given no weights are specified for any instances in the dataset.
>
> I've attempted to build the classifier through the pwk3 wrapper, the Java API, and the Weka GUI in Weka versions 3.8.6, 3.9.5, and 3.9.6 but have consistently gotten this error.
>
> I ~believe~ this exact dataset has been used to create a MISMO classifier in older versions of Weka (circa 2014), though I'm not sure which version that was in.
>
> Any insight into how to further debug this issue would be greatly appreciated.

If you look at the exception in the console, you get:

weka.core.WekaException:
weka.filters.unsupervised.attribute.PropositionalToMultiInstance: Some
instance weights are not equal to 1 and scheme does not implement the
WeightedInstancesHandler interface!

It's not generated by the MISMO classifier from the
multiInstanceLearning package:
https://weka.sourceforge.io/doc.packages/multiInstanceLearning/weka/classifiers/mi/MISMO.html

But the PropositionalToMultiInstance filter from the
multiInstanceFilters package (which is used internally by MISMO):
https://weka.sourceforge.io/doc.packages/multiInstanceFilters/weka/filters/unsupervised/attribute/PropositionalToMultiInstance.html

It's possible that some changes in core Weka since 2017-11-22 (when
the latest version of the multiInstanceLearning package was published)
now result in this error.

@Mark Hall
Maybe you can have a look at it when you come back from your summer break?

Cheers, Peter
-- 
Peter Reutemann
Dept. of Computer Science
University of Waikato, Hamilton, NZ
Mobile +64 22 190 2375
https://www.cs.waikato.ac.nz/~fracpete/
http://www.data-mining.co.nz/
_______________________________________________
Wekalist mailing list -- [email protected]
Send posts to [email protected]
To unsubscribe send an email to [email protected]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html