1 / 5
Splitting the raw query string into individual terms is ___.
-
-
-
-
Tokenisation breaks the query into tokens (words/subwords) that downstream stages can process.
2 / 5
Reducing 'running' and 'ran' to a common root is ___.
-
-
-
-
Stemming or lemmatisation normalises word forms so a search for one form matches documents using another.
3 / 5
Automatically fixing 'iphnoe' to 'iphone' is ___ correction.
-
-
-
-
Spell correction (typo tolerance) maps misspelled queries to likely intended terms so users still get results.
4 / 5
Treating 'laptop' and 'notebook' as equivalent uses ___ expansion.
-
-
-
-
Synonym expansion broadens a query to include equivalent terms, improving recall for varied phrasing.
5 / 5
Deciding whether a query is navigational, informational, or transactional is ___ classification.
-
-
-
-
Intent classification routes the query appropriately (e.g. show a product page vs articles), tailoring the experience to what the user wants.