Text Mining in Cybersecurity: Exploring Threats and Opportunities
, Babette J. Bakker 2, Erik Boertjes 3, Mike Wilmer 1, Stephan Raaijmakers 1,4 and Rick van der Kleij 5,6
Round 1
Reviewer 1 Report
see the enclosed file
Comments for author File: 
Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
In this paper, the authors have presented the Horizon scanner tool, which mines text data from scientific papers and websites. A small user study with 3 cyber experts is conducted. The results of the study were negative. The tool is currently not at the point where it can outperform using general search engines. My main concerns are that:
(1) The text mining methods seem a bit too simple.
(2) Authors seem to have omitted some details about their methods (or maybe there is some confusion because of the writing style). For example, they mention (line 161) "The processed data is then used to train semantic models ..." What are these semantic models?
On line 256, there is a gamma parameter that has not been defined yet. This whole paragraph sort of assumes that the reader knows the details of reference 41. Since this is an important section of the paper, the authors need to do a much better job of explaining.
In line 268 "per item" should be clarified.
The "Most Important Terms" section should describe any preprocessing steps performed and the relationship with Entities (if any).
(3) The small number of experts in the user study limits the usefulness of the results.
(4) The writing needs significant improvement
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The revised version is definitively better and the higher number of experts questioned (20 now) strengthens the overall results. Probably it would have been even better after a year of testing. Overall it is ready for publication in my opinion.
Author Response
The revised version is definitively better and the higher number of experts questioned (20 now) strengthens the overall results. Probably it would have been even better after a year of testing. Overall it is ready for publication in my opinion.
Thank you for this positive reaction. We checked the document again for spelling issues and made some minor changes in the discussion.
Reviewer 2 Report
In this version, the authors have added the missing details and clarified the confusing statements.
Presentation is improved from previous version. A few minor things remain such as "chapter 5" in the Introduction.
Author Response
In this version, the authors have added the missing details and clarified the confusing statements.
Thank you for this positive reaction!
Presentation is improved from previous version. A few minor things remain such as "chapter 5" in the Introduction.
Thank you. We checked the document again for spelling issues and made some minor changes in the discussion.
