[
https://issues.apache.org/jira/browse/MAHOUT-838?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Lance Norskog updated MAHOUT-838:
-
Attachment: MatrixWritable.java
ConfusionMatrix.java
Replace these two files,
[
https://issues.apache.org/jira/browse/MAHOUT-838?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Lance Norskog updated MAHOUT-838:
-
Attachment: MAHOUT-838_mini.patch
Make the confusion matrix writable to a file when testing
[
https://issues.apache.org/jira/browse/MAHOUT-838?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13133255#comment-13133255
]
Lance Norskog edited comment on MAHOUT-838 at 10/22/11 6:04 AM:
[
https://issues.apache.org/jira/browse/MAHOUT-847?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13133279#comment-13133279
]
Sean Owen commented on MAHOUT-847:
--
No, EuclideanDistanceMeasure is a distance measure
I don't understand the question.
What are the two vectors. Suppose you have x, an N dimensional vector and
\Omega, a random 1xN projection. \Omega x is a 1-dimensional vector.
Where is the second 1-d vector?
On Fri, Oct 21, 2011 at 8:23 PM, Lance Norskog goks...@gmail.com wrote:
More
IMHO, the confusion here comes from the point that there are some of random
projections, following this scheme, that preserves all pairwise geodesic and
euclidean distances between points on the original space into the new one
with high prob.
In fact, non-linear manifold learning algorithms aims
[
https://issues.apache.org/jira/browse/MAHOUT-849?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sean Owen resolved MAHOUT-849.
--
Resolution: Fixed
Fix Version/s: 0.6
Assignee: Sean Owen
Good eye. In fact there were
[
https://issues.apache.org/jira/browse/MAHOUT-845?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13133347#comment-13133347
]
Frank Scholten commented on MAHOUT-845:
---
Or we could add a getTerm(String[]
Bringing this to dev@, mid-thread, per Grant's suggestion. There was a
brief and fruitful thread on private@ to discuss project governance,
but the topic has shifted such that it's useful to just talk on dev@.
If I may paraphrase: I expressed concern about the sprawl of code and
algorithms, aging
[
https://issues.apache.org/jira/browse/MAHOUT-838?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13133355#comment-13133355
]
Sean Owen commented on MAHOUT-838:
--
Lance I tried just this, but I still get compile
On Oct 22, 2011, at 2:19 PM, Sean Owen wrote:
Bringing this to dev@, mid-thread, per Grant's suggestion. There was a
brief and fruitful thread on private@ to discuss project governance,
but the topic has shifted such that it's useful to just talk on dev@.
If I may paraphrase: I expressed
[
https://issues.apache.org/jira/browse/MAHOUT-849?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13133409#comment-13133409
]
Hudson commented on MAHOUT-849:
---
Integrated in Mahout-Quality # (See
Thanks! good thread.
On Sat, Oct 22, 2011 at 3:30 PM, Grant Ingersoll gsing...@apache.org wrote:
1. We aim for releases every 6 months or so
2. We make a best guess up front about what bug fixes will be in that
release, but we also will, obviously, bring in other fixes as they are
reported
When the board looks at the health of a community, one of the
questions it asks (or so I am told) is, 'Is the community responsive
to requests for assistance?'
Now, the board's bar here is quite low. I'm not trying to suggest for
a moment that Mahout is in any danger of attracting unfavorable
Drat: I wrote 'is necessarily a badge of shame' when I meant to write
'is not necessarily a badge of shame'.
[
https://issues.apache.org/jira/browse/MAHOUT-849?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13133493#comment-13133493
]
Lance Norskog commented on MAHOUT-849:
--
Error: wanted 43, received 43. This usually
Non-linear learning does, however, provide much more power than linear
schemes can.
The great virtue of linear schemes is when you have massive dimensionality.
On Sat, Oct 22, 2011 at 2:14 AM, Federico Castanedo castanedof...@gmail.com
wrote:
In fact, non-linear manifold learning algorithms
[
https://issues.apache.org/jira/browse/MAHOUT-849?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13133499#comment-13133499
]
Ted Dunning commented on MAHOUT-849:
It usually indicates a bug in the test, though.
I feel like I am most closely aligned with Grant. Very little to add.
Like it or not, Mahout is a library, not a coherent product such as hbase.
It's a collection of algorithms connectied together with some fairly thin
structure and persitence glue, but thenglue rarely can go much beyond that.
The debate above seems pretty complete.
What are positive actions that will make Mahout healthier?
Suggestions from debate:
* Automated patch testing. This would cure 'rotting patch' problem.
* Chivvying contributors for detailed notes.
* ?
Personal concepts:
* Regression suite with real data.
[
https://issues.apache.org/jira/browse/MAHOUT-524?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13133524#comment-13133524
]
Shannon Quinn commented on MAHOUT-524:
--
If there are two DLS.runJob() methods and the
No, I have two n-d vectors and want to measure their distance. Suppose I
project both via the same 1xN random projection matrix, and then find the
delta of the two values. Is this a valid distance? An approximate Manhattan
distance?
On Sat, Oct 22, 2011 at 1:46 AM, Ted Dunning
[
https://issues.apache.org/jira/browse/MAHOUT-849?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13133566#comment-13133566
]
Lance Norskog commented on MAHOUT-849:
--
My code attempted to multiply two MxN
It is a valid distance, but it will not necessarily accurately reflect the
original distance. You need to preserve roughly log N dimensions to get
something reasonably accurate relative to the original distance (with high
probability).
To the extent that your original vectors are colinear with
On Oct 22, 2011, at 6:41 PM, Sean Owen wrote:
Thanks! good thread.
On Sat, Oct 22, 2011 at 3:30 PM, Grant Ingersoll gsing...@apache.org wrote:
1. We aim for releases every 6 months or so
2. We make a best guess up front about what bug fixes will be in that
release, but we also will,
On Oct 22, 2011, at 7:34 PM, Benson Margulies wrote:
When the board looks at the health of a community, one of the
questions it asks (or so I am told) is, 'Is the community responsive
to requests for assistance?'
I think we are, but of course we could be better.
Now, the board's bar here
26 matches
Mail list logo