Interesting that they are making such heavy wind of their 'innovations' like
loop unrolling. I would be interested in how much mileage they will get out
of those tricks in more complicated code. I do not intend to disparage their
approach, but maintaining code with too many cute optimizations
+1 from me. Looks great!
Will do. My aplogies! -Greg
On Fri, Oct 7, 2011 at 3:55 AM, Luc Maisonobe luc.maison...@free.fr wrote:
Le 07/10/2011 07:21, gr...@apache.org a écrit :
Author: gregs
Date: Fri Oct 7 05:21:17 2011
New Revision: 1179935
URL:
Yes, this application would be similar to the ones you nixed. It would be a
factory returning an abstract class, with the appropriate methods throwing
exceptions (like getRank for the non-pivoting version-as Ted mentions in a
later entry in this blog).
As Ted asks, what is the issue with this?
If you really think about, all of the decomposition classes should be
handled by factories. The decompositions all seem to occur in the
constructor. Everything else is derived from those results, so one could
argue that the actual decomposition code could be written very procedurally
and put into
On Thu, Oct 6, 2011 at 9:13 PM, Phil Steitz phil.ste...@gmail.com wrote:
Lets a) stop top-posting (Gilles has asked politely a couple of times now)
and b) stay focused on solving the problems we actually have. We could
endlessly debate refactoring approaches.
Sorry about the top posting, I
Hi Gilles,
The class passes all current tests for QRDecomposition. Are you suggesting I
rip out the QRDecomposition from OLSMultipleRegression...etc and then make
sure there are no test failures? Or are you suggestion a new set of tests?
-Greg
On Wed, Oct 5, 2011 at 5:16 AM, Gilles Sadowski
Yes, in the interest of brevity I left out the fact that rank is calculated
as well. This is exactly Luc's implementation from Marquardt-Levenberg.
Is this what you are talking about with the cryptic 'all'?
On Wed, Oct 5, 2011 at 7:20 AM, Ted Dunning ted.dunn...@gmail.com wrote:
All.
On Wed,
class should also have a
getRank() method since it is after all 'rank-revealing' and in most (or
all?) cases it would be more efficient than
SingularValueDecomposition.getRank().
Chris.
On 5 October 2011 05:05, Greg Sterijevski gsterijev...@gmail.com wrote:
My apologies! Forgot to tag
with something that works)
3. Keep all three until the next major release, marking QRDecomposition (the
class) as deprecated.
-Greg
On Wed, Oct 5, 2011 at 9:39 AM, Gilles Sadowski
gil...@harfang.homelinux.org wrote:
On Wed, Oct 05, 2011 at 08:46:55AM -0500, Greg Sterijevski wrote:
Hi Gilles
I am sympathetic to this point of view [one class two algorithms-Ted's
point], but it seems like a step into the C world. Could we not accomplish
the same through some factory method? We can have two classes, but the user
might only get a reference to the abstract base. The other reason is that
that the user really doesn't see much or care about.
No need to hide it, but absolutely no reason to put it in their face,
either.
Pushing WIP code is always good.
On Wed, Oct 5, 2011 at 5:06 PM, Greg Sterijevski gsterijev...@gmail.com
wrote:
I am sympathetic to this point of view [one
Pardon my ignorance of Abelian Fields, but what would be a use of this set
of classes? Do they simplify some calculation or make some code faster? The
question is, are they numerical?
On Tue, Oct 4, 2011 at 2:33 PM, Luc Maisonobe luc.maison...@free.fr wrote:
Le 03/10/2011 19:50, Mikkel Meyer
Hello all,
A while back I was interested in being able to do pivoting qr decomposition.
I noticed that Chris Nix submitted a patch, but he indicated that he had
more work to do (testing and filling in functionality). In the discussion
around this, Luc suggested that I look at the QR decomposition
My apologies! Forgot to tag the subject line.
On Tue, Oct 4, 2011 at 8:35 PM, Greg Sterijevski gsterijev...@gmail.comwrote:
Hello all,
A while back I was interested in being able to do pivoting qr
decomposition. I noticed that Chris Nix submitted a patch, but he indicated
that he had more
I think Gilles is on the mark here.
+1 For the break up
+1 For organizing along conceptual lines.
-Greg
On Mon, Oct 3, 2011 at 3:59 PM, Gilles Sadowski
gil...@harfang.homelinux.org wrote:
On Mon, Oct 03, 2011 at 05:42:15PM +0200, Mikkel Meyer Andersen wrote:
2011/10/3 Phil Steitz
...@harfang.homelinux.org wrote:
On Fri, Sep 30, 2011 at 05:35:28PM -0500, Greg Sterijevski wrote:
Gilles,
I have attached a test harness which shows one spot where BOBYQA fails.
See the NonLinearConjugateGradient thread.
I don't know where you attached a unit test.
For new tests that exercise
01, 2011 at 08:28:52PM -0500, Greg Sterijevski wrote:
Per Phil's suggestion, they are now in
test/../../optimizers/NISTBatteryTest.java See JIRA: MATH-678 for further
info.
There currently isn't much info over there...
Also, the new NISTBatteryTest test class is quite unwieldy. The usual
Hello All,
I have been playing with the non linear optimizers in Commons for a few days
now. The context in which I am working is Probit/Logit/Tobit type limited
dependent variables models. These are well known problems. I found a very
well known example of a probit estimation (
But they all fail... except for verifying that my objective is correct at
the converged values.
On Fri, Sep 30, 2011 at 4:51 PM, Greg Sterijevski gsterijev...@gmail.comwrote:
Hello All,
I have been playing with the non linear optimizers in Commons for a few
days now. The context in which I
Gilles,
I have attached a test harness which shows one spot where BOBYQA fails.
See the NonLinearConjugateGradient thread.
Thanks,
-Greg
On Thu, Sep 29, 2011 at 6:06 PM, Gilles Sadowski
gil...@harfang.homelinux.org wrote:
Hi.
The work to be done on BOBYQAOptimizer is discussed in issue
I have added 4 more 'easy' cases from NIST's web site. I will check the test
in, omitting the @Test annotation.
On Fri, Sep 30, 2011 at 5:33 PM, Greg Sterijevski gsterijev...@gmail.comwrote:
But they all fail... except for verifying that my objective is correct at
the converged values
Hold off on 675. I owe you a couple of changes, related to naming and some
other issues we discussed on the ML.
-Greg
On Thu, Sep 29, 2011 at 1:44 PM, Luc Maisonobe luc.maison...@free.frwrote:
Le 29/09/2011 18:04, Gilles Sadowski a écrit :
Hello.
Referring to:
Gilles,
I have done what you suggested and notice that it (the optimizer) takes
humongous steps in the search for an upper bound. I tried making the initial
step size smaller, but to no avail.
The Powell optimizer flies very quickly to an optimum. Maybe there is some
misunderstanding in my usage
Okay. As per your other email, I was not aware that the code was in flux.
Thanks,
-Greg
On Thu, Sep 29, 2011 at 6:16 PM, Gilles Sadowski
gil...@harfang.homelinux.org wrote:
Hi.
I would like to add the following method to BOBYQA:
/**
*
* @return the number of
Hello All,
I am working on limited dependent variable regressions. I am testing part of
a logit regression routine. As part of the estimation technique I am need to
do some nonlinear optimization. Using the
NonLinearConjugateGradientOptimizer, I submit my problem, but get the
following exception:
Gilles,
You must be an android or Jedi master to spot something that tiny! ;-) Yes,
I am aware of the lowercase convention. I do intend to go through and clean
up not only that, but other ugliness in Miller... I wanted to stay as close
to the notation of the article as possible while I tested it.
Okay, my bad. I thought I had looked through the list of enums, but I
obviously was not as thorough as I should have been. Will make the changes
that you suggest.
Thank you,
-Greg
On Fri, Sep 23, 2011 at 5:51 AM, Gilles Sadowski
gil...@harfang.homelinux.org wrote:
Hello.
Author: gregs
Either is fine by me... ;) In case my opinion was solicited.
On Thu, Sep 22, 2011 at 5:02 AM, Gilles Sadowski
gil...@harfang.homelinux.org wrote:
As a side note, I notice now that NonMonotonousSequenceException
is misnamed. It should be NonMonotoneSequenceException. I think
it
, 2011 at 8:39 AM, Gilles Sadowski
gil...@harfang.homelinux.org wrote:
On Wed, Sep 21, 2011 at 08:31:00PM -0500, Greg Sterijevski wrote:
Any objections to fixing this?
Having a method
public static boolean isMonotone(double[] val,
OrderDirection dir
Hello All,
In MathUtils there exists the method:
public static boolean checkOrder(double[] val, OrderDirection dir,
boolean strict, boolean abort) {
...code omitted...
}
I would like to replace it with the method:
public static boolean
Meant to say add, not replace. My apologies. -Greg
On Wed, Sep 21, 2011 at 5:05 PM, Greg Sterijevski gsterijev...@gmail.comwrote:
Hello All,
In MathUtils there exists the method:
public static boolean checkOrder(double[] val, OrderDirection dir
-0500, Greg Sterijevski wrote:
Meant to say add, not replace. My apologies. -Greg
I like this better! ;-)
[But, still, please check the intended meaning of the first argument of
(sub-classes of) MathIllegalArgumentException.]
Gilles
at 05:17:59PM -0500, Greg Sterijevski wrote:
Meant to say add, not replace. My apologies. -Greg
I like this better! ;-)
[But, still, please check the intended meaning of the first argument of
(sub-classes of) MathIllegalArgumentException.]
Gilles
:
On 9/21/11 4:33 PM, Greg Sterijevski wrote:
Gilles,
I do not understand why a non-monotone collection should throw a
IllegalArgumentException...? There is nothing wrong with the argument, it
just is not in corrected order. Wouldn't it be better to return a false?
I think as you guys
...@gmail.com wrote:
On 9/21/11 4:33 PM, Greg Sterijevski wrote:
Gilles,
I do not understand why a non-monotone collection should throw a
IllegalArgumentException...? There is nothing wrong with the argument, it
just is not in corrected order. Wouldn't it be better to return a false?
I
Any objections to fixing this?
On Wed, Sep 21, 2011 at 8:27 PM, Phil Steitz phil.ste...@gmail.com wrote:
On 9/21/11 6:11 PM, Greg Sterijevski wrote:
One more question, there is a boolean argument called 'abort', what sense
does it make to keep checking an array given you have found one
should be thrown. I agree with
Gilles that it should be application specific (for linear iterative
solvers, the basic unit would probably be the number of matrix-vector
products). So I'll just avoid catching the MaxCountExceededException.
Best regards for now,
Sébastien
2011/9/11 Greg Sterijevski
the fix has been cut. So you sort of
jumped the gun on this one. OK (IMO) to leave this one as is, but
if we want to be tidy, we should reopen and the RESOLVE.
Phil
On 9/9/11 9:23 PM, greg sterijevski (JIRA) wrote:
[
https://issues.apache.org/jira/browse/MATH-649?page
Hi All,
Another mostly exceptional question from me! In the
interface UpdatingMultipleLinearRegression, we have regress() and
regress(int[] variablesToInclude).
1. What is the appropriate exception to throw when there is a variable index
in the array which does not exist in the data? For example
Sorry about the formatting. I looked through checkstyle report and my zeal
to eliminate extraneous whitespace was only matched by my need to add
whitespace around arguments!!!
On Sat, Sep 10, 2011 at 10:22 AM, Phil Steitz phil.ste...@gmail.com wrote:
On 9/9/11 9:18 PM, gr...@apache.org wrote:
Made the changes in my copy, will push them along with the changes to make
SimpleRegression compliant with the UpdatingMultipleRegression interface.
On Sat, Sep 10, 2011 at 12:29 PM, Greg Sterijevski
gsterijev...@gmail.comwrote:
Sorry about the formatting. I looked through checkstyle report
One more question, if the request list is null, what is the proper return? A
null or is an exception returned?
On Sat, Sep 10, 2011 at 12:15 PM, Greg Sterijevski
gsterijev...@gmail.comwrote:
Hi All,
Another mostly exceptional question from me! In the
interface
, 2011 at 1:20 PM, Gilles Sadowski
gil...@harfang.homelinux.org wrote:
On Sat, Sep 10, 2011 at 01:02:30PM -0500, Greg Sterijevski wrote:
One more question, if the request list is null, what is the proper
return? A
null or is an exception returned?
I think that we didn't quite decided
Thanks, -Greg
On Sat, Sep 10, 2011 at 4:11 PM, Gilles Sadowski
gil...@harfang.homelinux.org wrote:
Wrong ML thread.
On Sat, Sep 10, 2011 at 03:49:25PM -0500, Greg Sterijevski wrote:
Sebastian,
The only observation I made is in the JIRA notes. Use thread safe
collections
I sent the original to the wrong thread, my apologies!
Sebastian,
The only observation I made is in the JIRA notes. Use thread safe
collections for the listener lists. There is a bit of overhead, but if you
have long running optimizations you might want to connect, get the state of
the
Looks good to me.
On Fri, Sep 9, 2011 at 6:39 AM, Gilles Sadowski
gil...@harfang.homelinux.org wrote:
On Thu, Sep 08, 2011 at 04:55:46PM -0700, Ted Dunning wrote:
OK.
Replace that with the correct value. I meant it to be an index.
That doesn't change my other points. There is an
grandparent
is the Apache POM.
On 9 September 2011 06:15, Ralph Goers ralph.go...@dslextreme.com wrote:
FWIW, I built the commons vfs site with Maven 3 and didn't have any
issues.
On Sep 8, 2011, at 10:00 PM, Greg Sterijevski wrote:
Okay will do. My primary computer died, so I moved
wrote:
FWIW, I built the commons vfs site with Maven 3 and didn't have any
issues.
On Sep 8, 2011, at 10:00 PM, Greg Sterijevski wrote:
Okay will do. My primary computer died, so I moved to a Mac. Its a bit
of
learning curve, even coming from linux.
Thank you,
-Greg
On Thu
Added short comment. -Greg
On Tue, Aug 30, 2011 at 5:37 AM, sebb seb...@gmail.com wrote:
On 29 August 2011 07:11, Greg Sterijevski gsterijev...@gmail.com wrote:
My apologies, I was looking at the original c code and neglected to
notice
the cast.
Perhaps add a comment to the cast to make
Okay, I have begun nibbling on the doc. I will push a few changes I made in
connection to doing no intercept regression in SimpleRegression.
On Sun, Aug 28, 2011 at 10:17 PM, Phil Steitz phil.ste...@gmail.com wrote:
On 8/28/11 7:24 PM, Greg Sterijevski wrote:
Hello All,
What tool is used
or inverse condition number).
On Thu, Sep 8, 2011 at 2:41 AM, Phil Steitz phil.ste...@gmail.com wrote:
On 9/7/11 9:26 PM, Greg Sterijevski wrote:
I thought that QR is of O(n^3) complexity, while LU is probably in the
vicinity of O(n^2.5). While this is not a barn burning improvement
Cholesky, in my opinion, is not robust as you have discovered. When it
encounters a non-psd matrix it gives up. Maybe that is the correct course of
action, but I still think that when you are using the getCovariance to
estimate the curvature in the neighborhood of a point it would be okay to
take
...
On Thu, Sep 8, 2011 at 10:22 AM, Gilles Sadowski
gil...@harfang.homelinux.org wrote:
On Thu, Sep 08, 2011 at 09:49:12AM -0500, Greg Sterijevski wrote:
Cholesky, in my opinion, is not robust as you have discovered. When it
encounters a non-psd matrix it gives up. Maybe that is the correct course
Gilles,
Is the test checked in? I am interested in looking at it.
Thank you,
-Greg
On Thu, Sep 8, 2011 at 2:51 PM, Ted Dunning ted.dunn...@gmail.com wrote:
On Thu, Sep 8, 2011 at 12:42 PM, Luc Maisonobe luc.maison...@free.fr
wrote:
... Luc - are there other reasons that QR would be
I agree with Ted. You sometimes run into this issue with psuedoinverses. You
discover a zero eigenvalue at index i. The user's natural inclination is to
attribute it to the corresponding column of the data matrix.
-Greg
On Thu, Sep 8, 2011 at 6:55 PM, Ted Dunning ted.dunn...@gmail.com wrote:
Hello All,
I am seeing this. My build is on a Mac. Does anyone have any clues?
[ERROR] Failed to execute goal
org.apache.maven.plugins:maven-site-plugin:3.0-beta-3:site (default-site) on
project commons-math: failed to get Reports: Unable to parse configuration
of mojo
maven 2. That's what I use at least. You might try that.
Phil
On Sep 8, 2011, at 9:19 PM, Greg Sterijevski gsterijev...@gmail.com
wrote:
Hello All,
I am seeing this. My build is on a Mac. Does anyone have any clues?
[ERROR] Failed to execute goal
org.apache.maven.plugins:maven
It is also my recollection that LU is very quick to calculate. Would it be
possible to allow users to choose?
On Wed, Sep 7, 2011 at 3:07 PM, Ted Dunning ted.dunn...@gmail.com wrote:
Does the LUDecomposition not use pivots? LU should always do so since it
is
numerically unstable otherwise.
, Ted Dunning ted.dunn...@gmail.com wrote:
It shouldn't be all that different from QR (at most about 2x different).
On Wed, Sep 7, 2011 at 1:16 PM, Greg Sterijevski gsterijev...@gmail.com
wrote:
It is also my recollection that LU is very quick to calculate. Would it
be
possible to allow
Do you have unit tests written for the different components?
On Sat, Sep 3, 2011 at 10:48 AM, Patrick Meyer meyer...@gmail.com wrote:
Sounds good. Separating the general purpose stats from the psychometric
specific stats seems like a natural way to distinguish the two libraries.
I'll send a
/javadoc/com/kutsyy/util/package-summary.html
Are there plans to add a multivaraite normal CDF function to
commons math?
Patches welcome :)
Phil
Patrick
On 9/3/2011 12:36 PM, Greg Sterijevski wrote:
Do you have unit tests written for the different components?
On Sat, Sep 3, 2011 at 10
that the old SVD was ditched. Again, not saying it did
not deserve it, just wanted to make sure I understood the philosophy behind
the decision making.
Thank you,
-Greg
On Thu, Sep 1, 2011 at 1:11 AM, Phil Steitz phil.ste...@gmail.com wrote:
On 8/31/11 10:05 PM, Greg Sterijevski wrote:
Hello
Hello All,
This question popped into my head this evening, what is the right way to
handle multiple algorithms which purport to calculate the same thing? There
are, for example, a couple of ways to calculate the student t cdf. What is
the common's philosophy on deciding:
1. Whether to allow
Yes, I think this is a good idea. I will do it.
On Tue, Aug 30, 2011 at 5:37 AM, sebb seb...@gmail.com wrote:
On 29 August 2011 07:11, Greg Sterijevski gsterijev...@gmail.com wrote:
My apologies, I was looking at the original c code and neglected to
notice
the cast.
Perhaps add a comment
My apologies, I was looking at the original c code and neglected to notice
the cast.
On Mon, Aug 29, 2011 at 12:56 AM, Greg Sterijevski
gsterijev...@gmail.comwrote:
Yes, you seem to have a good point. My bad.
On Mon, Aug 29, 2011 at 12:50 AM, Ted Dunning ted.dunn...@gmail.comwrote:
Why
Hello All,
What tool is used to create the documentation (if one is used at all)?
Specifically, I am looking in commons-math/src/site/xdoc/userguide
If I wanted to add to the regression portion of stat.xml what is the best
way to accomplish this (use vi or some wysiwyg editor)?
Thank you,
Hello All,
In the MersenneTwister implementation in commons, I notice the following:
@Override
public void setSeed(int seed) {
// we use a long masked by 0xL as a poor man unsigned int
long longMT = seed;
mt[0]= (int) longMT;
for (mti = 1; mti N;
Yes, you seem to have a good point. My bad.
On Mon, Aug 29, 2011 at 12:50 AM, Ted Dunning ted.dunn...@gmail.com wrote:
Why?
Isn't that what casting as an (int) does?
On Sun, Aug 28, 2011 at 10:17 PM, Greg Sterijevski
gsterijev...@gmail.comwrote:
While probably not a big deal, shouldn't
Thank you David. In all honesty, either JCS or ehcache would be an
improvement over the my hackneyed implementation of a cache... ;-) -Greg
On Sat, Aug 27, 2011 at 9:14 AM, David Karlsen davidkarl...@gmail.comwrote:
I'd really recommend ehcache.
Den 27. aug. 2011 07:39 skrev Greg Sterijevski
Ted,
When you say
Functions are good, but giving a tiny bit
more information to the function is also a great idea
do you mean information on indexing and shape of the data?
One thought I had, I am not sure if this is 100% applicable is the
following:
1. You have two types of data (in matrix)
point, but, its an empirical question as to
whether my examples (as archetypes) are more likely to occur in practice
versus your example.
-Greg
On Fri, Aug 26, 2011 at 1:41 PM, Ted Dunning ted.dunn...@gmail.com wrote:
On Fri, Aug 26, 2011 at 7:38 AM, Greg Sterijevski gsterijev...@gmail.com
wrote
Hello,
I was poking around commons and noticed the dormant project Commons Cache.
The subversion history shows the last push occurred in 2007. What is the
story with the project? Has other java technology supplanted this project?
Thanks,
-Greg
, Aug 26, 2011 at 8:55 PM, Greg Sterijevski
gsterijev...@gmail.com wrote:
Hello,
I was poking around commons and noticed the dormant project Commons
Cache.
The subversion history shows the last push occurred in 2007. What is the
story with the project? Has other java technology supplanted
Before we fill JIRA with a bunch of point-counterpoint comments. Perhaps we
can organize our ideas on the list first?
Using Gille's organization:
* Consistency in naming and make explicit the rationale for choosing one or
another naming scheme
What methods are inconsistently named? Name the
.
On Wed, Aug 24, 2011 at 10:44 PM, Phil Steitz phil.ste...@gmail.com wrote:
On 8/24/11 7:27 PM, Greg Sterijevski wrote:
Before we fill JIRA with a bunch of point-counterpoint comments. Perhaps
we
can organize our ideas on the list first?
Using Gille's organization:
* Consistency in naming
Hello All,
Since math gives eigendecomposition for symmetrics, why even allocate the
array imagEigenvalues ?
What am I missing?
Thanks,
-Greg
I understand you want to support the general case, but why should it
necessitate instantiating a ref.
On Tue, Aug 23, 2011 at 5:29 PM, Luc Maisonobe luc.maison...@free.frwrote:
Le 23/08/2011 20:45, Greg Sterijevski a écrit :
Hello All,
Since math gives eigendecomposition for symmetrics
Why not use Ted's insight? Instead of a bunch of these setDiagonal,
setAll, setFirstN, ... why not just have a set( MatrixFunction mf ). The
MatrixFunction delegate could handle whatever we want. It would keep the
objects very clean. You could even have a class factory of commonly used
Should I open a ticket, or do you want to handle this?
On Tue, Aug 23, 2011 at 5:47 PM, Luc Maisonobe luc.maison...@free.frwrote:
Le 24/08/2011 00:44, Greg Sterijevski a écrit :
I understand you want to support the general case, but why should it
necessitate instantiating a ref.
You
Alright Phil! Thunder and Lightening! I like it. ;-)
On Mon, Aug 22, 2011 at 10:55 PM, Phil Steitz phil.ste...@gmail.com wrote:
On 8/22/11 8:42 PM, Greg Sterijevski wrote:
If no one has objections, I would like to harmonize simpleregression with
the Regression Interfaces. What is the best
Hello All,
In TriaDiagonalTransformer, I notice the following (commencing at line 104).
final double[] hK = householderVectors[k - 1];
final double inv = 1.0 / (secondary[k - 1] * hK[k]);
cachedQt.setEntry(k, k, 1);
if (hK[k] !=
If no one has objections, I would like to harmonize simpleregression with
the Regression Interfaces. What is the best way to proceed?
On Sun, Aug 21, 2011 at 9:25 PM, Greg Sterijevski gsterijev...@gmail.comwrote:
Opened a ticket. Submitted patches. -Greg
On Sat, Aug 20, 2011 at 4:47 PM, Phil
Opened a ticket. Submitted patches. -Greg
On Sat, Aug 20, 2011 at 4:47 PM, Phil Steitz phil.ste...@gmail.com wrote:
On 8/12/11 9:30 PM, Phil Steitz wrote:
On 8/12/11 7:16 PM, Greg Sterijevski wrote:
Hello All,
Before I chum the water with more JIRA tickets, I thought I would see
On symmetrics, diagonal, banded and so on, I disagree-as I have made clear
in the past. In the case of White standard errors or panel regressions, you
typically have long strings of multiplication by diagonals and symmetrics,
sandwich products and so forth. There are enough of these types of
I am building in Netbeans and I believe that Luc is correct. I initially had
problems until I included the junit.jar. (junit-4.8.2.jar)
PS What is your approach to the covariance matrix? Are you building a sum of
squares matrix or keeping updates of the mean, ... etc?
On Wed, Aug 17, 2011 at
Shocking as this may seem! ;-) I like Ted's suggestion. Very clean, an
appropriate use of OO and something for which the penalty is not great, but
the benefit humongous! +1 for functional/functor approach.
On Wed, Aug 17, 2011 at 6:44 PM, Ted Dunning ted.dunn...@gmail.com wrote:
Are you going
Forgive me for pushing my nose under the tent... I couldn't resist.
I think Gilles is saying that each specialization of the matrix/vector
objects would need to support pre (and post) multiplication with a dense. So
the type issue would not be problematic.
On Mon, Aug 15, 2011 at 6:34 PM, Ted
Also, I was thinking that maybe it might be useful to bring SimpleRegression
into line with the other regression techniques and give the user the ability
to constrain the constant to zero?
-Greg
On Sat, Aug 13, 2011 at 12:13 AM, Greg Sterijevski
gsterijev...@gmail.comwrote:
One more thing
10:59 PM, Greg Sterijevski wrote:
Also, I was thinking that maybe it might be useful to bring
SimpleRegression
into line with the other regression techniques and give the user the
ability
to constrain the constant to zero?
+1 - just make sure to include clear javadoc and tests.
Phil
Hello All,
Before I chum the water with more JIRA tickets, I thought I would see
whether people thought this was important.
In the SimpleRegression you have two methods:
public void addData(double x, double y) {
...some code that is not germane to discussion..
if (n 2) {
+1 from me, on all counts! -Greg
On Fri, Aug 12, 2011 at 11:30 PM, Phil Steitz phil.ste...@gmail.com wrote:
On 8/12/11 7:16 PM, Greg Sterijevski wrote:
Hello All,
Before I chum the water with more JIRA tickets, I thought I would see
whether people thought this was important
+= dy * dy * fact2;
sumXY += dx * dy * fact2;
xbar += dx * fact1;
ybar += dy * fact1;
I realize that most likely the compiler or runtime does this optimization,
but just in case...
-Greg
On Fri, Aug 12, 2011 at 11:40 PM, Greg Sterijevski
gsterijev
At least three with some code I checked in last night. The point is that
there is no reason to replicate the same thing over and over again.
-Greg
On Thu, Aug 11, 2011 at 6:33 AM, Gilles Sadowski
gil...@harfang.homelinux.org wrote:
In the SVD class I notice:
FastMath.max(m, n)
Gilles,
I think we do agree on most points:
For very-limited functionality, this could be an option. The big drawback
is that, if it takes time to implement, the merge with the official
branch
may not be trivial. [IIUC, Ted suggested that git was superior to svn
in making this less of a
Luc,
I think we misunderstood each other, in the snippet below, my intention was
to agree with him. Everything Gilles was agreeable and hard to argue
against.
-Greg
Luc
Not sure what you mean.
In my opinion, discussions should serve to solve real problems of a
software
library:
Hello All,
In the SVD class I notice:
FastMath.max(m, n)
all over the place. Since these values are known when the constructor is
called and are final, would anyone object to making the result a private
instance variable?
-Greg
I think you hit the nail on the head with the comment:
... but also the time to
experiment. Only the latter will be able to tell if the design is good.
And this must take time so that all the potential pitfalls can be ...
Maybe this is chumming the water with more flotsam and jetsam, but a lot
Does anyone have a list of showstopper bugs and features. If we want a 3.0
release soon, then perhaps some kind of a prioritized list would help us in
the community concentrate some of our efforts? I was thinking that the
individuals with commit privileges could put this together since they are
My bad! This was indeed for math commons...
On Tue, Aug 9, 2011 at 5:02 PM, Matt Benson gudnabr...@gmail.com wrote:
3.0 came out a couple of weeks ago. :) 3.0.1 is on the way now.
Matt
On Tue, Aug 9, 2011 at 4:59 PM, Greg Sterijevski gsterijev...@gmail.com
wrote:
Does anyone have
1 - 100 of 159 matches
Mail list logo
