On 05/17/2012 03:01 PM, Sean Owen wrote:
Yes I think you can and should certainly share a Recommender. It is
thread-safe. That wil in turn let you share data structures like
DataModel. Note you could run two different Recommenders from the same
shared DataModel if you wanted to, maybe to try two algorithms, but
that's not usual.
That's a separate question from how you handle similarities, and I
suggest you try caching in memory first to see if that gives you an
acceptable trade-off. You can pre-compute similarities if you want;
that's separate.
On Thu, May 17, 2012 at 12:57 PM, Nikolaos Romanos Katsipoulakis
<[email protected]> wrote:
On 05/17/2012 02:06 PM, Sean Owen wrote:
I think it will be way too slow to update a file like this, as you must
seek and rewrite the whole thing. Files aren't for this sort of update. A
DB might be.
Even if it were it is probably too slow to read the similarity from
backing
store each time. It needs to be mostly in memory. That's why the caching
wrapper is probably what you want.
Sean
On May 17, 2012 11:25 AM, "Nikolaos Romanos Katsipoulakis"<
[email protected]> wrote:
Hey guys,
In my recommender i need to save my UserSimilarity (CachingUserSimilarity
to be exact) in a file, and whenever I get a request for a
recommendation,
i need to load it from a file, retrieve the recommendation, and then
store
it back in a file. Can this be done? I have seen in the mahout
documentation a class FileItemSimilarity, which has a constructor that
parses the FileItemSimilarity from a file. Is there a similar
FileUserSimilarity? How is the FileItemSimilarity represented in the
file?
Thank you
Thank you for your immediate answer. Hmm, I think you are right. The problem
is that I want to maintain a static Recommender Object in a web service (and
use it as a singleton for all invocations of the web service), that is
updated asynchronously, and is persistent among all threads. Do you think
that this is a more acceptable design in matter of performance?
Thank you
That is wonderful. Thank you for your support. I really appreciate it.
Yours sincerely,
Nick Katsipoulakis.
