On 06/24/2017 11:23 PM, Shujing Ke wrote:
Yes, to output the pattern. There are two ways:
1. return the atomspace that store the patterns.
2. return a HandleSeq of patterns.
As far as the C++ API is concerned a HandleSeq, or an
OrderedHandleSet/UnorderedHandleSet depending on whether we want to
retain the order or not. Likewise for the scheme interface, a ListLink
or SetLink.
Each raw pattern will be quoted within a PatternLink.
Oh you mean
PatternLink
QuoteLink
<pattern>
?
Why not. That means that the postprocessing to turn patterns into useful
atoms would occur outside of the pattern miner. Maybe we should do that
for now, then once the post-processing is well understood it can be
moved inside to the pattern miner.
Then we still need to represent the pattern scores, I've never
experimented with proto-atoms but I think that would be the way to go,
either that or using Evaluation or Execution, such as
Execution <1 1>
GroundedSchema "pattern-count"
List
PatternLink
...
Number "42"
telling there are 42 instances of that pattern. I suppose it is called
frequency in the context of the pattern miner (don't understand why it,
BTW).
Because the numbers of patterns are huge, I think it make sense to give
an input parameter to specify the top percentage of patterns to be
output. e.g.: only output top 10% frequent patterns; output top 7%
frequent and 20% interesting patterns. Or directly specify the number
of frequency : like frequency > 15
I suppose for now the frequency (what I remember is called the support,
right?) would be enough. As it would control computational effort as well.
My baby is still sick, having serious diarrhea, does not want to have
much food, need to go to hospital from time to time. So I probably won't
get much work done this week. Hope he will be better next week.
Sorry to hear to about that, I know this can be very distressful.
Nil
Shujing
On Wed, Jun 21, 2017 at 12:45 PM, Ben Goertzel <[email protected]
<mailto:[email protected]>> wrote:
On Wed, Jun 21, 2017 at 1:50 PM, Nil Geisweiller
<[email protected] <mailto:[email protected]>> wrote:
> OK. Although I think we still need to come up soon with a way to pass the
> results, including frequencies, interestingness, etc, as atoms in the
> atomspace, as opposed to writing the results in a file.
Agreed...
--
Ben Goertzel, PhD
http://goertzel.org
"I am God! I am nothing, I'm play, I am freedom, I am life. I am the
boundary, I am the peak." -- Alexander Scriabin
--
You received this message because you are subscribed to the Google Groups
"opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/opencog.
To view this discussion on the web visit
https://groups.google.com/d/msgid/opencog/c81475c0-9aa1-b673-6409-9ab3ae306710%40gmail.com.
For more options, visit https://groups.google.com/d/optout.
