Hi Everyone!
Working on a project that requires a Search query similiar to what is
seen on "amazon.com" in that after searching for and displaying an
item, the system shows:
"Users that have searched for "AAAAA" AND "BBBBB" have also searched
for "CCCC".
Where "BBBBB" and "CCCC" are other related entities.
1. Has anyone used or heard of this algorithm?
2. Is Lucene capable of executing "learning searches"?
3. Is Lucene capable of executing suggested Searches?
Any assistance rendered would be appreciated.
Thanks in advance.
javaman
Patricia Shanahan - 07 Jun 2006 00:51 GMT
> Hi Everyone!
>
[quoted text clipped - 16 lines]
>
> javaman
The general subject of using observed behavior of a group of users to
predict behavior of an individual is called "collaborative filtering".
There are many algorithms - see
http://citeseer.ist.psu.edu/marlin04collaborative.html for a fairly
technical, reasonably recent, survey paper.
There is a paper, "Amazon.com Recommendations Item-to-Item Collaborative
Filtering" by Greg Linden, Brent Smith, and Jeremy York of Amazon.com.
I found it on-line at:
http://ieeexplore.ieee.org/iel5/4236/26323/01167344.pdf?tp=&arnumber=1167344&isn
umber=26323
but I don't know whether you will be able to access it at that URL.
Patricia
Chris Uppal - 07 Jun 2006 09:34 GMT
> Working on a project that requires a Search query similiar to what is
> seen on "amazon.com" in that after searching for and displaying an
> item, the system shows:
>
> "Users that have searched for "AAAAA" AND "BBBBB" have also searched
> for "CCCC".
If a user searches for AAAA one day, then BBBB two months later, should the
system suggest AAAA to a different user looking for BBBB ?
In either case, I would imagine that if you stored previous queries in a Lucene
DB (either grouped by user or not), then you could use a normal Lucene search
to find previous related queries.
Note that if you do store queries, and especially if you store queries
associated with individual users, then there might be serious privacy concerns.
-- chris