Agustin Mauro asked 3 years ago
Hi!
I’m trying to build a list of authors for a 40 years period articles corpus.
For term extraction there is an option to extract terms dynamically choosing how many time slices there are. There is no similar option for list building. I imagine that if I try to build the list without the time dependent part then recent authors will be over-represented in the list because they have more articles. I could split the corpus in two, build two separated list and join them, but I imagine there is an easier option.
Regards
Agustin
Lionel Staff answered 3 years ago
Dear Augustin,
True, there is not way to build lists using list builder script based on some pre-defined periods.
We have just deploy a new feature associated with Epic Epoch script: it is now able to produce a csv file which correspond to the bump graph. So:
- Define the number of top entity you want per period
- Select the way you want to build your periods (period slicer or period detector or directly defined using years)
- Download the csv file, and build a cross table in order to have author names in line and period in column
- Duplicate the first column with the author names in the second column, work on the second column to harmonized the names if needed
- Index back you dataset using list indexer script
I hope it helps
L