CIS server linguistic analysis module uses stemming to recognize grammatical forms of a word and treat them as a single evidence term. Stemming means extracting the common root, or stem, from words. For example, the words parked, parks, and parking share the same stem (park). The CIS server recognizes them as four instances of the same evidence term rather than as four different terms.
Turn off stemming when a common noun is used as a proper noun. For example, if you define the term Explorer as in Microsoft Internet Explorer, you do not want CIS server to recognize other forms of the word as the same term. Also turn off stemming to treat different forms of the same stem as separate terms; for example, to use provider and provide as evidence of different categories.
When you turn off stemming, CIS server looks only for an exact match of the defined term. Explicitly add as terms all of the forms you want CIS server to recognize, such as plural forms or different forms of the verb.
You can activate the stemming at different levels:
In the category class definition, you can use the stemming on the category names. If you select Use stemming in the category class definition, then it is the default value for all categories created from this category class.
In the category definition, you can override the option inherited from the category class to use the stemming on the category names.
For each evidence term, you can use the stemming, unless you selected Any language as the category language. In this case, the option is disabled.
Related topics: