In the writing of a research paper, I've collected a large quantity of public blog posts and ingested them into a LDA and a word2vec data model. When I'm using terms from within these models, how can I reference the data source both to make it clear that this is analysis of my corpus but also to avoid plagiarism ambiguity?

As I don't know the specific location of these terms within the blogs (because the terms I'm dealing with are coming OUT of the data models) I cant cite the specific article.

What's the general consensus on citing a self-created corpus, particularly if the terms and language has been disconnected from their original articles.

  • 2
    I would publish the corpus as a separate data entity, probably as a supplementary material to your paper. You seem not to have a copyright problem, do you? Commented Feb 1, 2018 at 11:39
  • @OlegLobachev, Thank you for your advice. I had planned on making the corpus and data models available on a github, referencing the terms and phrases produced from the data models from that would clear up the confusion? No, there's no copyright issue. Thank you for your help
    – Dove
    Commented Feb 2, 2018 at 7:07
  • 1
    You are welcome. If the volume is high and no changes are in sight, a data repository like Zenodo might be more fitting the purpose. Commented Feb 2, 2018 at 14:59


You must log in to answer this question.

Browse other questions tagged .