Fuzzy Matching and Stemming

Fuzzy matching enables you to match similarly spelled words in queries. Oracle Text provides entity extraction for multiple languages.

Stemming enables you to match words with the same linguistic root. For example a query on $speak, expands to search for all documents that contain speak, speaks, spoke, and spoken.

Fuzzy matching and stemming are automatically enabled in your index if Oracle Text supports this feature for your language.

Fuzzy matching is enabled with default parameters for its fuzzy score and for its maximum number of expanded terms. Fuzzy score is a measure of how closely the expanded word matches the query word. At index time, you can change these default parameters.

To automatically detect the language of a document and to have the necessary transformations performed, create a stem index by enabling the index_stems attribute of the AUTO_LEXER. Use the stemmer that corresponds to the document language and always configure the stemmer to maximize document recall. For compound words in languages such as German, Finnish, Swedish, and Dutch, if you set index_stems to YES, then compound word stemming is automatically performed in the documents. Compounds are always separated into their component stems.

To improve the performance of stem queries, create a stem index by enabling the index_stems attribute of BASIC_LEXER.

• Values For Language Attribute for index_stems of AUTO_LEXER

• Values For Language Attribute for index_stems of BASIC_LEXER

See Also: Oracle Text Reference to learn more about fuzzy matching and stemming