STEMMING

In the English language, the root or underlying form of a word is called a stem. Stems serve as the basis for different variations of a word. For example, the word sell has the variants seller, selling, and sellable. The term stemming refers to the ability of Internet search engines to search for all of a word's possible variants to return more comprehensive search results. Therefore, someone searching for information about selling on the Internet could simply type in sell (or one of its variants, like seller) and the engine would return results with all possible variants of the word sell. This is obviously a great help for consumers and business professionals who conduct Internet research related to e-commerce.

Some Internet search engines, as well as the search tools and information retrieval systems on databases and other computer systems, allow users to enable or disable stemming. The manner by which this is achieved varies depending on how a search tool operates. On some applications it might involve checking a box or placing a character like the plus sign next to a word to enable the function. Conversely, a minus sign after a word, or some other character, might be used to disable the function. Stemming also can be used selectively, meaning that if someone were looking for information about selling apples, stemming could be enabled for the word selling, but not for the word apple. Therefore, irrelevant variations of the word apple, such as applet and applecart, would be excluded from the search results.

Stemming often is used as an alternative to wild-cards. Wildcards are symbols that can be substituted for a character or value during an information search. By including a dollar sign after the keyword sell (sell$), a search engine might look for Internet resources containing different variations of that word (sell, selling, and seller), working in a similar fashion to stemming.

FURTHER READING:

"Stem Classes." The Center for Intelligent Information Retrieval. July 20, 2001. Available from ciir.cs.umass.edu/cgi-bin/stemming.

SEE ALSO: Search Engine Strategy

stemming The returning of documents which contain words that are prefixed by a search term provided by the user of a SEARCH ENGINE. For example, a search engine which uses stemming responds to a query containing the keyword ‘run’ with all words starting with ‘run’, such as ‘running’ and ‘runner’.