Stemming involves reducing query terms to their most basic form, or stem, and using this to identify all the derivatives of that word such a plurals which should be considered a match.
For example, searching for
initialization
will also return results for initialize, initials, etc. Results matching the exact query term searched initialization will rank higher than its derivatives in the search results list.
Stemming can be disabled by using double quotes around words or phrases. Thus it is possible to search for a precise variant of a word including plurals or singular variants of words, for example
“dog” NOT dogs.
Word Stemming is enabled by default for all publication text fields, such as full-text and abstract.
Note that wildcards are not automatically stemmed.