Suffix Processing

kzinda · Post by **kzinda** » Thu Mar 17, 2005 9:10 am

When searching our patent database for
"inhaler dispense w/2",

the engine doesn't find

"dispensed for oral inhalation".

The engine requires an asterisk in the search string:
"inhaler dispens* w/2"

Why doesn't the suffix processing routine identify "dispensed" from "dispense"?

Post by **mark** » Thu Mar 17, 2005 9:57 am

What's your minwordlen setting? By default it's 255 effectively disabling suffix processing.

And unless you've "set withinmode='word'" w/2 means within 2 characters.

kzinda · Post by **kzinda** » Thu Mar 17, 2005 10:50 am

minwordlen = 5, withinword=word

Also note that the engine finds "inhalation" from "inhaler" indicating suffix processing is on. The fact that it finds the phrase with the asterisk indicates word proximity.

Post by **mark** » Thu Mar 17, 2005 2:05 pm

dispense is one of those odd words all forms of which don't strip to the same root using the default suffixes.

dispense doesn't strip down, but dispen[s][ed] does.