Stop Word

Definition

Stop Word

A stop word is a word type on a chosen stop list, whose occurrences are excluded from indexing or query processing. Stop words are usually frequent words judged to contribute little to distinguishing relevant documents, not words without meaning.

Filtering Trade-off

Removing common words such as the, a, and, to and be can reduce index size and query-processing work. Compression and query optimisation reduce the cost of retaining them, making removal less attractive.

Filtering can also discard information needed for exact phrase matching, titles or relationships expressed by a query. The choice of stop list therefore depends on the language and retrieval task.

Example

Example

• king of Denmark: of is part of the phrase being sought.
• Let it be and To be or not to be: removing common words can erase most or all of a title or quotation.
• flights to London: removing to loses the distinction between London as a destination and other relationships, such as flights from London.