Replies: 1 comment
-
Of course! Here's your initial response with grammar corrections: This is an excellent idea! I have seen people re-importing a dataset to continue screening with a different model or to re-label the initially excluded papers to double-check their relevance, but this is yet another great idea! If you export the file with labels and re-import this file, ASReview will recognize the labels and use these to train the first iteration of the model. If you want to re-label these records, you need to remove the labels or rename the tag we use. You could export only the relevant records (since this pull request was merged) as a .xlsx or .csv file (instead of .ris) and simply rename the column used for the labels. In this table, you can see which labels we use, so you need to choose a different column name. If you then import this file into ASReview, it will treat all records as unseen. Make sure to also add a few irrelevant records to help with the training stage. If you want to convert your file back to .ris, you can use convert in Datatools. |
Beta Was this translation helpful? Give feedback.
-
Hi! just a quick question,
before i begin i want to say thank you! this tool is amazing !!
i'm working on a scope review whereby i've collected citations from several databases, compiled them, removed duplicates and now after assessing most of them I found 700, some what relevant, references which i now want to submit to a second team. The question is, can the second team take the 700 relevant citations and run them again through as-review as they have a more clear inclusion/exclusion criteria?
I've already tired to test this by exporting the relevant citations and setting up a new project with those papers, but every-time i do that, it shows all the citations as labelled, specifically "relevant". how do i make the AI forget!?
UPDATE i realise it's because when i export the relevant citations theres a note on each citation which says "AS-REVIEW_relevant". So i suppose my question is can i export it without the note on each citation?
Beta Was this translation helpful? Give feedback.
All reactions