> How does the GSA decide what query suggestion is most relevant? For example
> if a Query Suggestion is clicked on 50 time over one that has been used 2
> times which is displayed first?
My guess is that it's the frequency of searches. Honestly, I've never
looked, and don't have an easy way to test this right now, so I'll
have to wait and see. But only successful searches are used to build
the query suggestion list if I recall correctly - successful meaning
that it returns results, not that it returns the desired results of
course.
> If the relevance built in the GSA advanced learning? If yes can we bias the suggestions?
No, I don't think the GSA has a lot of "learning" involved with query
suggestion, and I'm not aware of any way to bias them.
> Also how does the GSA handle something like a swear word in a phrase. Does it blacklist the phrase "your
> a %^&*" as I suspect it would?
I don't think there's any blacklisting done by the query suggestion
mechanism itself, and I don't think the GSA does anything to handle
profanity. But unless you have documents that contain profanity, those
searches probably won't return results, unless you have other
non-profane keywords within the query that are within your documents I
guess.
> Sorry; one last question how does the
> appliance handle words that are improper but someone has substituted a
> special character or number to resemble a letter?
I don't think it does anything to handle them, but they won't return
results unless you have documents that contain those words.