Custom synonym dictionaries for enterprise search collections

After you create an enterprise search collection, you can add synonym dictionaries to the system and allow users to expand the search results by searching for documents that contain synonyms of the query terms.

Restriction: If you create a content analytics collection, you can define equivalent terms for words when you create a custom user dictionary. When users explore the collection with the content analytics miner, documents that contain the equivalent terms are included in the analysis results. You cannot create a synonym dictionary to associate with a content analytics collection.

Synonyms typically include multiple word terms such as product names like WebSphere® Application Server. Multiple word terms that are contained in the synonym dictionary are correctly identified in user queries and do not have to appear within quotes.

The Search and Index API (SIAPI) supports several ways for users to search for synonyms of query terms when searching collections:

The SIAPI query syntax supports the tilde (~) operator for synonym expansion. If the user prepends this operator to a query term, synonym expansion is performed for the word. For example, the query ~WAS returns documents that discuss WebSphere Application Server and any other synonyms that exist for this abbreviation.
Synonym expansion can be enabled using the SIAPI synonym expansion interface from within a search application. Query terms can be automatically expanded to include synonyms, or the search application might include options that enable the user to specify whether synonyms of the query terms are to be returned in the search results.
During automatic synonym expansion, synonym lookup is performed on all query words. The search results include documents that contain either the query terms or synonyms of the query terms. The SIAPI also supports the generation of a list of synonym expansions for the submitted query.
Synonym expansion in n-gram collections allows for phrase segmentation of the query text. If an entire phrase appears in the synonym dictionary, then the search succeeds. A phrase is extracted according to these delimiters:
Punctuation

The following characters are delimiters: - ( ) + . ,
Quotation marks are ignored and do not delimit phrases.

Change in alphabet

For example, for an n-gram collection, the query will be expanded to include the synonyms of ABC in the following sample queries if ABC is in the synonym dictionary:
ABC run DEF stand (where ABC and DEF are Japanese text)
ABC+DCF+GHI

You must define synonyms in an XML file which you must then convert to a synonym dictionary so that it can be added to the system and associated with a collection.

You can select which synonym dictionary to use in the administration console. One synonym dictionary can be selected for each enterprise search collection. A synonym dictionary can be shared by several collections.