We used the following keywords files to identify categories for journals and conferences not in Scopus, for our STI 2023 paper "Assessing the agreement in retraction indexing across 4 multidisciplinary sources: Crossref, Retraction Watch, Scopus, and Web of Science".

The first four text files each contains keywords/content words in the form: 'keyword1', 'keyword2', 'keyword3', .... The file title indicates the name of the category:
file1: healthscience_words.txt
file2: lifescience_words.txt
file3: physicalscience_words.txt
file4: socialscience_words.txt

The first four files were generated from a combination of software and manual review in an iterative process in which we:
- Manually reviewed venue titles were not able to automatically categorize using the Scopus categorization or extending it as a resource.
- Iteratively reviewed uncategorized venue titles to manually curate additional keywords as content words indicating a venue title could be classified in the category healthscience, lifescience, physicalscience, or socialscience. We used English content words and added words we could automatically translate to identify content words. NOTE: Terminology with multiple potential meanings or contain non-English words that did not yield useful automatic translations e.g., (e.g., Al-Masāq) were not selected as content words.

The fifth text file is a list of stopwords in the form: 'stopword1', 'stopword2, 'stopword3', ...
file5: stopwords.txt
This file contains manually curated stopwords from venue titles to handle non-content words like 'conference' and 'journal,' etc.

This dataset is a revision of the following dataset:
Version 1: Lee, Jou; Schneider, Jodi: Keywords for manual field assignment for Assessing the agreement in retraction indexing across 4 multidisciplinary sources: Crossref, Retraction Watch, Scopus, and Web of Science. University of Illinois at Urbana-Champaign Data Bank.

Changes from Version 1 to Version 2:
- Added one author
- Added a stopwords file that was used in our data preprocessing.
- Thoroughly reviewed each of the 4 keywords lists. In particular, we added UTF-8 terminology, removed some non-content words and misclassified content words, and extensively reviewed non-English keywords.