Stem search

[wsbones]

I did a search for 'marry' using the stemming option. I expected it would find marriage but it didn't. Should it? How does the option actually work?

Thanks.

 

[Brandon Staggs]

It uses an algorithm that breaks down a word to it's "stem," but it is not 100% perfect. It's good at finding stems from words like loving, loveth, etc (down to love). Here is a very technical explanation: http://www.tartarus.org/~martin/PorterStemmer/http://snowball.tartarus.org/algorithms/english/stemmer.html

 

[wsbones]

Thanks. I looked at the VB code. Here are the six steps in the algorithm in case anyone is interested:
1. Get rid of plurals
2. Turn "y" to "i" if a vowel exists in the stem
3. Map double suffixes to single suffixes
4. Deal with "ic", "ful", "ness", etc.
5. Remove "ant", "ence", etc.
6. Remove final "e"

The stem search results should be useful for developing a more refined search strategy.