"word vector with additional attributes"
dali
New Altair Community Member
Hi,
I'd like to use 1 CSV-file for classification. In this file there are several rows (ID, title, text, class). Now I want to build the word vector of the text (that works) and combine it with the weighted title. Something like 1/3 text and 2/3 title. Afterwards this should be used for classification, while the classes are also given in this one CSV-file.
· So how can I combine these two rows and weight them? (I also don't understand why its not enough to say these rows are of type text, but in addition having to convert them from nominal to text)
· How can I use the class as the label? (Giving the special attribute label seems not to be enough, have to set role.)
I'd be very glad, if you could give me a hint, maybe naming the operators I should try or giving the order of operations.
THX a lot,
dali.
I'd like to use 1 CSV-file for classification. In this file there are several rows (ID, title, text, class). Now I want to build the word vector of the text (that works) and combine it with the weighted title. Something like 1/3 text and 2/3 title. Afterwards this should be used for classification, while the classes are also given in this one CSV-file.
· So how can I combine these two rows and weight them? (I also don't understand why its not enough to say these rows are of type text, but in addition having to convert them from nominal to text)
· How can I use the class as the label? (Giving the special attribute label seems not to be enough, have to set role.)
I'd be very glad, if you could give me a hint, maybe naming the operators I should try or giving the order of operations.
THX a lot,
dali.
Tagged:
0
Answers
-
use the set role operator to set the class attribute as a "label" type for prediction
use the process documents from data operator and set the "specify weights" checkbox
i have a video series on text mining with RM here: http://www.youtube.com/user/VancouverData0