... maybe we should change the function to take a list of subject URIs
...
Agreed
--
apriori: get most used (websites/notes/documents/etc...)
https://bugs.launchpad.net/bugs/494288
You received this bug notification because you are a member of Zeitgeist
Developers, which is the registrant for
I have a little prototype of a full apriori! sadly it is very intensive
calculation. Basically I create a daily neighborhood over the last 3 weeks.
The results hve ot be tested.
As for our current apriori we need an adaptive way of generating our
neighborhoods instead of
pulling out only x
Okay, this is now merged into lp:zeitgeist (after doing some more
changes and having Markus review it again), but of course this doesn't
mean that we can't continue improving it there (all the opposite!).
You are right. I think a hybrid approach would give the best results
Yeah. But for this we
** Branch linked: lp:zeitgeist
--
apriori: get most used (websites/notes/documents/etc...)
https://bugs.launchpad.net/bugs/494288
You received this bug notification because you are a member of Zeitgeist
Developers, which is the registrant for Zeitgeist Framework.
Status in Zeitgeist Framework:
The biggest problem right now is i only consider the first 7 items after
document X as part of the shopping list.
We have 3 options to optimize our results.
- We pick out the all entires between the open and close events of X
- We pick out all entries between an open and a modfiy event. If an open
2009/12/9 Mikkel Kamstrup Erlandsen mikkel.kamst...@gmail.com
I don't think we should consider open/close events when calculating
these relations. That way it wont work for contacts and other non-file-
like items.
The initial step of the algorithm: Fetch the last 7 events for this
subject
Siegfried: Ouch. Got me :-) How about if we do ORDER count(subj_uri)
LIMIT 5 instead? It would change the ordering (quelle surprise :-)), but
I am wondering what the correct ordering here is? By popularity or by
recency?
--
apriori: get most used (websites/notes/documents/etc...)
Seif: Fx. delta = 15 minutes, (=90ms). So you take into account all
events that happened in a 15 minute radius of the timestamp.
As far as I can see the way you do it would be meen that you consider
two events related if there are no other events in between?
I honestly don't know which
8 matches
Mail list logo
