Chapter 7. Search Methodology

Table of Contents

Stemming
Order of Results

Stemming

A search for dogs will also return hits with the word dog and a search for parenting will return results with the words parent and parental. This is because the search uses stemming to help return the most relevant results. That is, words are reduced to their stem (or root word) before the search is performed.

The stemming algorithm relies on common English language patterns - like verbs ending in ing - to find the stems. This is more efficient than looking up each search term in a dictionary and usually produces desirable results. However, it also means the search will sometimes reduce a word to an incorrect stem and cause unexpected results. To prevent a word or phrase from stemming, put it in double-quotes to force an exact search. For example, a search for parenting will also return results for parental, but a search for "parenting" will not.

Understanding how stemming works can help you to create more relevant searches, but it is usually best not to anticipate how a search term will be stemmed. For example, searching for gold compass does not return the same results as golden compass, because -en is not a regular suffix in English, and therefore the stemming algorithm does not recognize gold as a stem of golden.