I've installed weka and the python-weka-wrapper.
I got as far as
from weka.classifiers import Classifier
clf=Classifier(classname="weka.classifiers.rules.JRip")
from random import randint
X = [[randint(1,10) for _ in range(5)] for _ in range(100)]
y = [randint(0,1) for _ in range(100)]
but now I don't know how to load my data which is available as a Python data structure.
How can I load my data matrices, output the rules (in some parsable format) and test the classifier on new data?
You can create data on the fly, by defining the structure (ie attributes) and then adding the data rows. Also, I added a convenience method to instantiate Instances to the weka.core.dataset module called create_Instances_from_lists, which is scheduled for the next release.
This question has been answered on the project's mailing list as well, including examples and pointers.