Folks,
In my experience, you can save a good deal of memory and some time by
implementing a 2d matrix as a single vector and some multiplication, instead
of the [][] resulting in a vector of pointers to vectors. Would a patch for
this purpose be viewed as helpful, or was my experience anomalous?
I've looked at the implementation of Canopy in DisplayKMeans, and then tried
to compare it to the MapReduce version.
I'm sure that the simple version in DisplayKMeans has the potential to loop
indefinitely. I can't prove that the problem exists in the real map/reduce
code one way or the other.
Th
This is a great thing to have.
In general, I think that rows and columns of matrices should be labeled. A
vector is either a row or a column and thus it should have a label and its
elements should have labels. It should not cost you anything if you don't
want labels and I do.
Tracing back to a
I think I have a grip on this, but I'm not quite sure. Please forgive me if
I'm confused and this belongs back on 'users'.
Drivers like the K-Means driver eat vectors and emit clusters of vectors.
The vectors don't carry any sort of field that can be used to back-associate
them with documents or o
[
https://issues.apache.org/jira/browse/MAHOUT-129?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714632#action_12714632
]
Jeff Eastman commented on MAHOUT-129:
-
The KMeansDriver numCentroids argument is incorr
[
https://issues.apache.org/jira/browse/MAHOUT-129?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jeff Eastman resolved MAHOUT-129.
-
Resolution: Fixed
Rename completed
> Kmeans sample does not expose numIterations control from KM
if it is Class, we need then an instance of Class> and I
don't know how to get one.
On May 29, 2009, at 3:36 PM, Ted Dunning wrote:
Probably needs Class
WHat mumble should be is tricky to say without looking at the code.
Object
is a candidate.
On Fri, May 29, 2009 at 6:16 AM, Sean Owen
[
https://issues.apache.org/jira/browse/MAHOUT-127?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sean Owen updated MAHOUT-127:
-
Priority: Minor (was: Major)
Affects Version/s: 0.2
Fix Version/s: 0.2
[
https://issues.apache.org/jira/browse/MAHOUT-127?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sean Owen updated MAHOUT-127:
-
Attachment: warnings.patch
How about this patch? removed the serialVersionUID and tweaked what appeared t
[
https://issues.apache.org/jira/browse/MAHOUT-127?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714604#action_12714604
]
Benson Margulies commented on MAHOUT-127:
-
OK, I'm convinced that I should turn off
I can't find any documentation that 1L is a special value, that
prompts the JVM to go construct its own value. I find some posts to
the contrary too, like
http://www.nabble.com/serialVersionUID-td23001300.html and
https://jira.jboss.org/jira/browse/SECURITY-341
Setting it to 1L is worse than setti
"?" and "? extends Object" should be equivalent and yeah nothing seems
to work for any value of mumble.
On Fri, May 29, 2009 at 8:36 PM, Ted Dunning wrote:
> Probably needs Class
Probably needs Class
WHat mumble should be is tricky to say without looking at the code. Object
is a candidate.
On Fri, May 29, 2009 at 6:16 AM, Sean Owen wrote:
> Anyway, yeah if you make the type parameter Class, then it
> complains that, essentially, Class.class is not of this type, which
>
The problem with the ant stuff is that it is a lot of work for an end-user
to copy and incorporate outside of the mahout tree. So, to make the example
more exemplary, shade is attractive.
On Fri, May 29, 2009 at 12:56 PM, Grant Ingersoll wrote:
> Feel free to submit a patch. The ant approach wor
[
https://issues.apache.org/jira/browse/MAHOUT-128?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Benson Margulies updated MAHOUT-128:
Attachment: pom.diff
Here's the right patch.
> maven parent not included in build
> --
[
https://issues.apache.org/jira/browse/MAHOUT-128?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Benson Margulies updated MAHOUT-128:
Attachment: (was: pom.diff)
> maven parent not included in build
>
[
https://issues.apache.org/jira/browse/MAHOUT-129?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Benson Margulies updated MAHOUT-129:
Component/s: (was: Collaborative Filtering)
Clustering
Summary:
[
https://issues.apache.org/jira/browse/MAHOUT-129?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Benson Margulies updated MAHOUT-129:
Description:
The KMeans driver forces the numReduceTasks parameter of KMeans to 1, and ther
[
https://issues.apache.org/jira/browse/MAHOUT-129?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714518#action_12714518
]
Benson Margulies commented on MAHOUT-129:
-
This patch isn't quite right, as follows
[
https://issues.apache.org/jira/browse/MAHOUT-126?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714515#action_12714515
]
Grant Ingersoll commented on MAHOUT-126:
Shashikant,
Couple of comments on the Luc
[
https://issues.apache.org/jira/browse/MAHOUT-129?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Benson Margulies updated MAHOUT-129:
Attachment: kmeans-patch.diff
> kmeans sample makes one cluster
> -
kmeans sample makes one cluster
---
Key: MAHOUT-129
URL: https://issues.apache.org/jira/browse/MAHOUT-129
Project: Mahout
Issue Type: Bug
Components: Collaborative Filtering
Affects Versions: 0.2
[
https://issues.apache.org/jira/browse/MAHOUT-128?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714510#action_12714510
]
Benson Margulies commented on MAHOUT-128:
-
yes the problem is that the pom isn't in
Feel free to submit a patch. The ant approach works for now, but I'm
open to better ways of doing it.
On May 29, 2009, at 12:01 PM, Benson Margulies wrote:
http://maven.apache.org/plugins/maven-shade-plugin/
I post an example of this once I get through the puzzle of how to
map input
to a
[
https://issues.apache.org/jira/browse/MAHOUT-126?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714509#action_12714509
]
Grant Ingersoll commented on MAHOUT-126:
So just kind of brainstorming here, but I
[
https://issues.apache.org/jira/browse/MAHOUT-128?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714499#action_12714499
]
Grant Ingersoll commented on MAHOUT-128:
I'm not following what this does. When I
http://maven.apache.org/plugins/maven-shade-plugin/
I post an example of this once I get through the puzzle of how to map input
to a SparseVector instead of a vector.
On Fri, May 29, 2009 at 11:57 AM, Grant Ingersoll wrote:
> Link?
>
>
> On May 29, 2009, at 11:51 AM, Benson Margulies wrote:
>
>
Link?
On May 29, 2009, at 11:51 AM, Benson Margulies wrote:
Any reason not to use maven-shade-plugin to make jobs?
Any reason not to use maven-shade-plugin to make jobs?
[
https://issues.apache.org/jira/browse/MAHOUT-128?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Benson Margulies updated MAHOUT-128:
Attachment: pom.diff
> maven parent not included in build
> ---
maven parent not included in build
--
Key: MAHOUT-128
URL: https://issues.apache.org/jira/browse/MAHOUT-128
Project: Mahout
Issue Type: Bug
Reporter: Benson Margulies
The maven parent isn't inclu
[
https://issues.apache.org/jira/browse/MAHOUT-127?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714487#action_12714487
]
Benson Margulies commented on MAHOUT-127:
-
Are you sure? Eclipse claims that coding
By default ID do you mean no serialVersionUID? Then I agree. Setting any
particular value though would introduce the concern I mentioned. And fixing
it to a particular vale like 1 would be even more problematic.
On May 29, 2009 4:11 PM, "Benson Margulies (JIRA)" wrote:
[
https://issues.apach
[
https://issues.apache.org/jira/browse/MAHOUT-127?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Benson Margulies updated MAHOUT-127:
Attachment: warnings.diff
Here is a patch with less gratuitous changes. Perhaps few enough?
[
https://issues.apache.org/jira/browse/MAHOUT-127?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Benson Margulies updated MAHOUT-127:
Attachment: (was: warnings.diff)
> Remove warnings
>
>
>
[
https://issues.apache.org/jira/browse/MAHOUT-127?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714480#action_12714480
]
Benson Margulies commented on MAHOUT-127:
-
The cast changes were overuse of the Ecl
[
https://issues.apache.org/jira/browse/MAHOUT-127?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714477#action_12714477
]
Sean Owen commented on MAHOUT-127:
--
- I like the changes you mentioned, like @SuppressWarn
[
https://issues.apache.org/jira/browse/MAHOUT-127?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Benson Margulies updated MAHOUT-127:
Attachment: warnings.diff
> Remove warnings
>
>
> Key: MA
Remove warnings
Key: MAHOUT-127
URL: https://issues.apache.org/jira/browse/MAHOUT-127
Project: Mahout
Issue Type: Bug
Reporter: Benson Margulies
The patch I'm about to attach gets rid of all the current yellow
Ah right. I am not sure why I don't see a warning on that anymore in IntelliJ 9.
Anyway, yeah if you make the type parameter Class, then it
complains that, essentially, Class.class is not of this type, which
doesn't make sense to me.
I agree, it is worth @SuppressWarnings. Send over your patch an
Eclipse always warns for a naked use of Class. So, Eclipse thinks, more or
less, that any time you might want to say X, that you really meant to
say X>, or X>, which is in some cases the same
sort of beast. However, this is not one of those cases. I think that an
@SuppressWarnings is called for.
O
I don't think so... that class doesn't need a new parameter. It is an
AbstractParameter, parameterized by Class, already.
Actually, I opened this up again in IntelliJ and no longer see the
unchecked-cast sort of warnings I remember on this code. I certainly
recall something like that here.
What's
Should that class really be ClassParameter instead of just
ClassParameter?
On Fri, May 29, 2009 at 8:13 AM, Sean Owen wrote:
> I personally would be very into it. I use IntelliJ and yeah it yells
> about a whole lot of the same stuff every time I open a file.
>
> I also cannot figure out what to
[
https://issues.apache.org/jira/browse/MAHOUT-126?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Benson Margulies updated MAHOUT-126:
Attachment: mahout-126-benson.patch
Improved patch. Allows specification of file character
[
https://issues.apache.org/jira/browse/MAHOUT-126?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714426#action_12714426
]
Benson Margulies commented on MAHOUT-126:
-
This patch needs to explicitly manage th
I personally would be very into it. I use IntelliJ and yeah it yells
about a whole lot of the same stuff every time I open a file.
I also cannot figure out what to do with that Class line!
On Fri, May 29, 2009 at 1:11 PM, Benson Margulies wrote:
> Can I interest you in a patch to zero out the ec
Can I interest you in a patch to zero out the eclipse 3.4 warning count?
This would involve deleting some unused variables, removing some
unnecesssary @SuppressWarning's, and in one case (ClassParameter) adding an
@SupressWarning, since I can't for the life me figure out how to make that
into
clas
[
https://issues.apache.org/jira/browse/MAHOUT-126?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714414#action_12714414
]
Grant Ingersoll commented on MAHOUT-126:
See SOLR-1193.
> Prepare document vecto
[
https://issues.apache.org/jira/browse/MAHOUT-126?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Grant Ingersoll reassigned MAHOUT-126:
--
Assignee: Grant Ingersoll
> Prepare document vectors from the text
> --
[
https://issues.apache.org/jira/browse/MAHOUT-126?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714401#action_12714401
]
Grant Ingersoll commented on MAHOUT-126:
Passing in a way to make a custom weight o
[
https://issues.apache.org/jira/browse/MAHOUT-126?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714362#action_12714362
]
David Hall commented on MAHOUT-126:
---
Sure, I just want to be able to have:
double weig
[
https://issues.apache.org/jira/browse/MAHOUT-126?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714356#action_12714356
]
Shashikant Kore commented on MAHOUT-126:
David,
Sorry, I don't have any background
[
https://issues.apache.org/jira/browse/MAHOUT-126?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Shashikant Kore updated MAHOUT-126:
---
Attachment: MAHOUT-126.patch
Patch to create index and document vectors from text.
> Prepare
[
https://issues.apache.org/jira/browse/MAHOUT-126?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714348#action_12714348
]
David Hall commented on MAHOUT-126:
---
I actually need something like this as well for LDA,
Prepare document vectors from the text
--
Key: MAHOUT-126
URL: https://issues.apache.org/jira/browse/MAHOUT-126
Project: Mahout
Issue Type: New Feature
Reporter: Shashikant Kore
Clustering al
55 matches
Mail list logo
