Hi All,
I would like to calculate the sum of all values in a row in a dataset while the number of columns can vary. It is a text mining case so that it is not possible to predict the exact number of columns generated by vector creation. I would like to generate a new attribute containing the sum value. the dataset could look like
term; text1; text2; text3
free; 2;3;1
mind; 4;2;8
suggest;3;1;1
telephone;1;0;6
the result should be
term; text1; text2; text3;sum of occ
free;2;3;1;6
mind;4;2;8;14
suggest;3;1;1;5
telephone;1;0;6;7
as I said, the count of texts can vary so as far as I understand it the Generate Attributes operater cannot do the job.
Thanks in advance!
Chris
The generate aggregation operator should be fine for this.
Set the attribute filter type parameter to "value_type", the value type parameter to "numeric" and the aggregation function to "sum" in order to make it only work on numeric attributes.
regards
Andrew