Re: Filter dataset by column value Wrapper Python

> I would like to do this but using Wrapper.
>
> data_frame = pd.read_csv ('positions.csv')
> vehs= pd.unique(data_frame['id'])
>
> for i in range(0, len(vehs)):
>   df0 = data_frame[data_frame['id']==vehs[i]]
>   dat = np.array(df0[['latitude','longitude']])
>
> Is it possible to do it using Wrapper on the fly? I mean without saving them in parallel in CSV, convert them to arff and load them.

As long as your ID is a NOMINAL or STRING attribute, then yes (you can
use the NumericToNominal filter to turn a numeric ID attribute into a
nominal one).

For simplicity, the following example uses the class attribute of the
iris UCI dataset as ID for the subsets generated by the
SubsetByExpression filter (ATTx with x being 1-based index):

import weka.core.jvm as jvm
from weka.core.converters import load_any_file
from weka.filters import Filter

jvm.start()

data = load_any_file("/some/where/iris.arff")
vals = data.attribute(data.num_attributes - 1).values
for val in vals:
    print("\n\n", "-->", val, "\n")
    # generate subset
    f= Filter(classname="weka.filters.unsupervised.instance.SubsetByExpression",
           options=["-E", "ATT5 is '%s'" % val])
    f.inputformat(data)
    subdata = f.filter(data)
    # create subset of columns (you can also use col_range="1,2")
    subdata = subdata.subset(col_names=["sepallength", "sepalwidth"])
    print(subdata)  # here you would do something with the data

jvm.stop()

Cheers, Peter
-- 
Peter Reutemann
Dept. of Computer Science
University of Waikato, NZ
+64 (7) 858-5174 (office)
+64 (7) 577-5304 (home office)
http://www.cms.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