September 2022 data-update for "Updated science-wide author databases of standardized citation indicators"

28 October 2022 update note:
The previous September 2022 update has now been amended to account for a limited field assignment in the release of October 10, 2022 (https://elsevier.digitalcommonsdata.com/datasets/btchxktzyw/4).
Version 5, in line with previous versions, assigns authors to 174 scientific sub-fields. Version 5, with respect to version 4, has a higher agreement of author sub-field assignments with the data from previous year releases (versions 1-3). The Science Metrix classification of scientific fields and subfields has been used for all releases of the database to-date. However, at the time of the release of version 4, Science-Metrix classification was happening largely based on machine learning assignment of articles into scientific fields and subfields and this works better for some fields than for others. As a result, some subfields (e.g. Architecture or Epidemiology) had few or no scientists assigned to them. Version 5 reinstated a hybrid approach where articles are assigned first to a scientific field and subfield based on the journal of their publication. For multidisciplinary journals then the machine learning approach is used for assigning articles that have no journal based assignment. For ranking purposes, each scientist is then assigned to the subfield where he/she has published the largest number of articles. The database shows the first and second subfield for each scientist.  
No metric values have changed between version 4 and version 5, for instance all citation scores are identical. However, since some scientists have now been assigned to a different subfield in version 5 than in version 4, it is possible that they may be included in the top-cited scientists in version 5 but not in version 4 (if they are in the top-2% based on their version 5 subfield classification but were not in the top-2% of their subfield based on the version 4 classification); or, vice versa, they may not be included in the top-2% based on their version 5 classification but might have been in the top-2% based on their version 4 classification. Version 5 is considered more appropriate and should replace version 4.