[jira] Commented: (MATH-512) GaussianFitter could inherit from CurveFitter
[
https://issues.apache.org/jira/browse/MATH-512?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=12996336#comment-12996336
]
Gilles commented on MATH-512:
-
{quote}
I already submitted the second patch earlier today.
{quote}
I know, and I looked at it; but it still contains many formatting errors (tab
characters, trailing white space, { characters on their own line).
The first two kinds of errors are detected by CheckStyle; you can generate
the report by running the following command:
{noformat}
$ mvn site
{noformat}
and the html report will be located in the file
{noformat}
target/site/checkstyle.html
{noformat}
For the third kind, you'll probably have to review the code manually.
I think that it is better to correct as many problems as possible before
committing any changes. Even so, some will remain, but the fewer, the better,
especially for those who clean-up afterwards ;-).
[By the way, I think that the name of the inner class
GaussianParametersGuesser could be shortened to maybe ParameterGuesser.]
GaussianFitter could inherit from CurveFitter
-
Key: MATH-512
URL: https://issues.apache.org/jira/browse/MATH-512
Project: Commons Math
Issue Type: Improvement
Reporter: Gilles
Priority: Minor
Fix For: 3.0
Attachments: GaussianFitter.patch, GaussianFitter.patch,
GaussianFitterTest.patch, ParametricGaussianFunctionTest.patch
Currently, the GaussianFitter class contains a CurveFitter instance (and
delegates most method calls to it) for the sake of being able to return a
GaussianFunction instance (from the call to the fit method).
I think that it would clearer (in the context of the fitting package) that
fit returns an array of doubles (as CurveFitter does). The user can then
easily create a Gaussian function object, if needed.
[Also, note that the GaussianFitter actually fits the sum of a constant and
a Gaussian function; thus the name is slightly misleading IMO.]
I also think that ParametricGaussianFunction and
GaussianParametersGuesser should become inner classes of GaussianFitter
(as this makes it clear what their use is).
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Commented: (MATH-512) GaussianFitter could inherit from CurveFitter
[ https://issues.apache.org/jira/browse/MATH-512?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=12996337#comment-12996337 ] Gilles commented on MATH-512: - I'd also prefer to wait for the approval concerning the dropping of the constant... GaussianFitter could inherit from CurveFitter - Key: MATH-512 URL: https://issues.apache.org/jira/browse/MATH-512 Project: Commons Math Issue Type: Improvement Reporter: Gilles Priority: Minor Fix For: 3.0 Attachments: GaussianFitter.patch, GaussianFitter.patch, GaussianFitterTest.patch, ParametricGaussianFunctionTest.patch Currently, the GaussianFitter class contains a CurveFitter instance (and delegates most method calls to it) for the sake of being able to return a GaussianFunction instance (from the call to the fit method). I think that it would clearer (in the context of the fitting package) that fit returns an array of doubles (as CurveFitter does). The user can then easily create a Gaussian function object, if needed. [Also, note that the GaussianFitter actually fits the sum of a constant and a Gaussian function; thus the name is slightly misleading IMO.] I also think that ParametricGaussianFunction and GaussianParametersGuesser should become inner classes of GaussianFitter (as this makes it clear what their use is). -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Commented: (MATH-512) GaussianFitter could inherit from CurveFitter
[ https://issues.apache.org/jira/browse/MATH-512?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=12996349#comment-12996349 ] Luc Maisonobe commented on MATH-512: I would also prefer to have drop the additive constant and have the gaussian function really be gaussian. It is always possible for users to build their own function with an additive constant (or a polynomial or anything else) in their own classes, delegating some of the comutation to this class and adding their own parameters. GaussianFitter could inherit from CurveFitter - Key: MATH-512 URL: https://issues.apache.org/jira/browse/MATH-512 Project: Commons Math Issue Type: Improvement Reporter: Gilles Priority: Minor Fix For: 3.0 Attachments: GaussianFitter.patch, GaussianFitter.patch, GaussianFitterTest.patch, ParametricGaussianFunctionTest.patch Currently, the GaussianFitter class contains a CurveFitter instance (and delegates most method calls to it) for the sake of being able to return a GaussianFunction instance (from the call to the fit method). I think that it would clearer (in the context of the fitting package) that fit returns an array of doubles (as CurveFitter does). The user can then easily create a Gaussian function object, if needed. [Also, note that the GaussianFitter actually fits the sum of a constant and a Gaussian function; thus the name is slightly misleading IMO.] I also think that ParametricGaussianFunction and GaussianParametersGuesser should become inner classes of GaussianFitter (as this makes it clear what their use is). -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Commented: (JEXL-110) conceptual improvement for the language
[ https://issues.apache.org/jira/browse/JEXL-110?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=12996380#comment-12996380 ] Henri Biestro commented on JEXL-110: JEXL is not intended as a replacement for ECMAScript; if you need richer constructs or semantics than what JEXL provides, you'd probably be off using scripting (aka JSR-223). If you can describe a strong use-case for the feature, I'll gladly look at it. Cheers conceptual improvement for the language --- Key: JEXL-110 URL: https://issues.apache.org/jira/browse/JEXL-110 Project: Commons JEXL Issue Type: Improvement Reporter: Cristian Lorenzetto the language dont distinguish when a object is null and when is undefined we could add a keyword undefined -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Resolved: (JEXL-108) parsing error if i define a empty literal array/map
[
https://issues.apache.org/jira/browse/JEXL-108?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Henri Biestro resolved JEXL-108.
Resolution: Fixed
Fix Version/s: 2.0.2
Assignee: Henri Biestro
Modified parser and associated classes to accept empty array literal [] and
empty map literal { : }.
Committed revision 1071986.
M src/test/java/org/apache/commons/jexl2/IssuesTest.java
M src/main/java/org/apache/commons/jexl2/parser/Parser.jjt
A src/main/java/org/apache/commons/jexl2/parser/ASTMapLiteral.java
M src/main/java/org/apache/commons/jexl2/parser/ASTArrayLiteral.java
M src/main/java/org/apache/commons/jexl2/Debugger.java
M pom.xml
parsing error if i define a empty literal array/map
---
Key: JEXL-108
URL: https://issues.apache.org/jira/browse/JEXL-108
Project: Commons JEXL
Issue Type: Bug
Affects Versions: 2.0.1
Reporter: Cristian Lorenzetto
Assignee: Henri Biestro
Fix For: 2.0.2
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Commented: (JEXL-83) Make JexlArithmetic immutable (and threadsafe)
[ https://issues.apache.org/jira/browse/JEXL-83?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=12996417#comment-12996417 ] Henri Biestro commented on JEXL-83: --- After several months of usage, the only use-case where the leniency has a useful reason to change (besides tests) can not be an engine wide setting. In the encountered cases, the engine was used both for configuration (in strict mode at startup) and during runtime evaluations (lenient). To avoid multiple Jexl engine instances - heavy -, the easy route was to derive JexlArithmetic so it could allow a thread-local behavior wrt leniency. This will be incorporated in the trunk. Make JexlArithmetic immutable (and threadsafe) -- Key: JEXL-83 URL: https://issues.apache.org/jira/browse/JEXL-83 Project: Commons JEXL Issue Type: Improvement Reporter: Sebb Assignee: Henri Biestro Fix For: Later Attachments: JEXL-83.patch The JexlArithmetic class is currently not thread-safe because of the single field: /** Whether this JexlArithmetic instance behaves in strict or lenient mode. */ protected boolean strict; If this field were made final, and the setLenient() method removed, then the class would be immutable and therefore thread-safe. This would mean that JexlEngine would also have to change. The class is new to 2.0, so this is the best time to fix this. -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Resolved: (JEXL-83) Make JexlArithmetic immutable (and threadsafe)
[ https://issues.apache.org/jira/browse/JEXL-83?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Henri Biestro resolved JEXL-83. --- Resolution: Fixed Fix Version/s: (was: Later) 2.0.2 Added JexlThreadedArithmetic that allows different threads to manipulate the arithmetic leniency per thread. Committed revision 1072000. M src/test/java/org/apache/commons/jexl2/CacheTest.java M src/test/java/org/apache/commons/jexl2/PublicFieldsTest.java M src/test/java/org/apache/commons/jexl2/UnifiedJEXLTest.java M src/test/java/org/apache/commons/jexl2/junit/Asserter.java M src/test/java/org/apache/commons/jexl2/IssuesTest.java M src/test/java/org/apache/commons/jexl2/AssignTest.java M src/test/java/org/apache/commons/jexl2/ArithmeticTest.java M src/test/java/org/apache/commons/jexl2/JexlTestCase.java M src/test/java/org/apache/commons/jexl2/IfTest.java M src/main/java/org/apache/commons/jexl2/JexlArithmetic.java A src/main/java/org/apache/commons/jexl2/JexlThreadedArithmetic.java M src/main/java/org/apache/commons/jexl2/JexlEngine.java M src/site/xdoc/changes.xml Make JexlArithmetic immutable (and threadsafe) -- Key: JEXL-83 URL: https://issues.apache.org/jira/browse/JEXL-83 Project: Commons JEXL Issue Type: Improvement Reporter: Sebb Assignee: Henri Biestro Fix For: 2.0.2 Attachments: JEXL-83.patch The JexlArithmetic class is currently not thread-safe because of the single field: /** Whether this JexlArithmetic instance behaves in strict or lenient mode. */ protected boolean strict; If this field were made final, and the setLenient() method removed, then the class would be immutable and therefore thread-safe. This would mean that JexlEngine would also have to change. The class is new to 2.0, so this is the best time to fix this. -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Resolved: (JEXL-103) Inputstream reads no byte if it is called into a class runnable
[ https://issues.apache.org/jira/browse/JEXL-103?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Henri Biestro resolved JEXL-103. Resolution: Cannot Reproduce Fix Version/s: 2.0.2 Assignee: Henri Biestro No test case and no news for a long time; dismissed as can not reproduce. Inputstream reads no byte if it is called into a class runnable Key: JEXL-103 URL: https://issues.apache.org/jira/browse/JEXL-103 Project: Commons JEXL Issue Type: Bug Reporter: Cristian Lorenzetto Assignee: Henri Biestro Fix For: 2.0.2 Inputstream instance reads no byte if it is called into a class runnable. if i call the same function in the java code it works , if i call it in the jexl code it enter in the function but than it dont read any bytes in the stream -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Updated: (MATH-512) GaussianFitter could inherit from CurveFitter
[ https://issues.apache.org/jira/browse/MATH-512?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Ole Ersoy updated MATH-512: --- Attachment: GaussianFitter.patch GaussianFitter could inherit from CurveFitter - Key: MATH-512 URL: https://issues.apache.org/jira/browse/MATH-512 Project: Commons Math Issue Type: Improvement Reporter: Gilles Priority: Minor Fix For: 3.0 Attachments: GaussianFitter.patch, GaussianFitter.patch, GaussianFitter.patch, GaussianFitterTest.patch, ParametricGaussianFunctionTest.patch Currently, the GaussianFitter class contains a CurveFitter instance (and delegates most method calls to it) for the sake of being able to return a GaussianFunction instance (from the call to the fit method). I think that it would clearer (in the context of the fitting package) that fit returns an array of doubles (as CurveFitter does). The user can then easily create a Gaussian function object, if needed. [Also, note that the GaussianFitter actually fits the sum of a constant and a Gaussian function; thus the name is slightly misleading IMO.] I also think that ParametricGaussianFunction and GaussianParametersGuesser should become inner classes of GaussianFitter (as this makes it clear what their use is). -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Commented: (MATH-512) GaussianFitter could inherit from CurveFitter
[ https://issues.apache.org/jira/browse/MATH-512?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=12996477#comment-12996477 ] Ole Ersoy commented on MATH-512: OK - I cleaned up the formatting. It still says there's an illegal tab on line 80. Is there a commons math formatting profile that I should be using, or should I just create a new one that only uses spaces instead of tabs? GaussianFitter could inherit from CurveFitter - Key: MATH-512 URL: https://issues.apache.org/jira/browse/MATH-512 Project: Commons Math Issue Type: Improvement Reporter: Gilles Priority: Minor Fix For: 3.0 Attachments: GaussianFitter.patch, GaussianFitter.patch, GaussianFitter.patch, GaussianFitterTest.patch, ParametricGaussianFunctionTest.patch Currently, the GaussianFitter class contains a CurveFitter instance (and delegates most method calls to it) for the sake of being able to return a GaussianFunction instance (from the call to the fit method). I think that it would clearer (in the context of the fitting package) that fit returns an array of doubles (as CurveFitter does). The user can then easily create a Gaussian function object, if needed. [Also, note that the GaussianFitter actually fits the sum of a constant and a Gaussian function; thus the name is slightly misleading IMO.] I also think that ParametricGaussianFunction and GaussianParametersGuesser should become inner classes of GaussianFitter (as this makes it clear what their use is). -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Commented: (MATH-512) GaussianFitter could inherit from CurveFitter
[ https://issues.apache.org/jira/browse/MATH-512?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=12996505#comment-12996505 ] Gilles commented on MATH-512: - I did some more formatting clean-up. Also removed a trivial method that was used only once. I don't know about profiles (I use emacs with some configuration that automatically indents the code with the appropriate amount of space characters). GaussianFitter could inherit from CurveFitter - Key: MATH-512 URL: https://issues.apache.org/jira/browse/MATH-512 Project: Commons Math Issue Type: Improvement Reporter: Gilles Priority: Minor Fix For: 3.0 Attachments: GaussianFitter.patch, GaussianFitter.patch, GaussianFitter.patch, GaussianFitterTest.patch, ParametricGaussianFunctionTest.patch Currently, the GaussianFitter class contains a CurveFitter instance (and delegates most method calls to it) for the sake of being able to return a GaussianFunction instance (from the call to the fit method). I think that it would clearer (in the context of the fitting package) that fit returns an array of doubles (as CurveFitter does). The user can then easily create a Gaussian function object, if needed. [Also, note that the GaussianFitter actually fits the sum of a constant and a Gaussian function; thus the name is slightly misleading IMO.] I also think that ParametricGaussianFunction and GaussianParametersGuesser should become inner classes of GaussianFitter (as this makes it clear what their use is). -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Resolved: (MATH-512) GaussianFitter could inherit from CurveFitter
[ https://issues.apache.org/jira/browse/MATH-512?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Gilles resolved MATH-512. - Resolution: Fixed Changes committed in revision 1072056. Thanks for the patch, Ole. GaussianFitter could inherit from CurveFitter - Key: MATH-512 URL: https://issues.apache.org/jira/browse/MATH-512 Project: Commons Math Issue Type: Improvement Reporter: Gilles Priority: Minor Fix For: 3.0 Attachments: GaussianFitter.patch, GaussianFitter.patch, GaussianFitter.patch, GaussianFitterTest.patch, ParametricGaussianFunctionTest.patch Currently, the GaussianFitter class contains a CurveFitter instance (and delegates most method calls to it) for the sake of being able to return a GaussianFunction instance (from the call to the fit method). I think that it would clearer (in the context of the fitting package) that fit returns an array of doubles (as CurveFitter does). The user can then easily create a Gaussian function object, if needed. [Also, note that the GaussianFitter actually fits the sum of a constant and a Gaussian function; thus the name is slightly misleading IMO.] I also think that ParametricGaussianFunction and GaussianParametersGuesser should become inner classes of GaussianFitter (as this makes it clear what their use is). -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Commented: (MATH-512) GaussianFitter could inherit from CurveFitter
[ https://issues.apache.org/jira/browse/MATH-512?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=12996522#comment-12996522 ] Luc Maisonobe commented on MATH-512: If you work with eclipse, you can use the following profile: http://people.apache.org/~luc/Apache-commons.xml GaussianFitter could inherit from CurveFitter - Key: MATH-512 URL: https://issues.apache.org/jira/browse/MATH-512 Project: Commons Math Issue Type: Improvement Reporter: Gilles Priority: Minor Fix For: 3.0 Attachments: GaussianFitter.patch, GaussianFitter.patch, GaussianFitter.patch, GaussianFitterTest.patch, ParametricGaussianFunctionTest.patch Currently, the GaussianFitter class contains a CurveFitter instance (and delegates most method calls to it) for the sake of being able to return a GaussianFunction instance (from the call to the fit method). I think that it would clearer (in the context of the fitting package) that fit returns an array of doubles (as CurveFitter does). The user can then easily create a Gaussian function object, if needed. [Also, note that the GaussianFitter actually fits the sum of a constant and a Gaussian function; thus the name is slightly misleading IMO.] I also think that ParametricGaussianFunction and GaussianParametersGuesser should become inner classes of GaussianFitter (as this makes it clear what their use is). -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Commented: (MATH-512) GaussianFitter could inherit from CurveFitter
[ https://issues.apache.org/jira/browse/MATH-512?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=12996524#comment-12996524 ] Ole Ersoy commented on MATH-512: Sweet - Thanks! GaussianFitter could inherit from CurveFitter - Key: MATH-512 URL: https://issues.apache.org/jira/browse/MATH-512 Project: Commons Math Issue Type: Improvement Reporter: Gilles Priority: Minor Fix For: 3.0 Attachments: GaussianFitter.patch, GaussianFitter.patch, GaussianFitter.patch, GaussianFitterTest.patch, ParametricGaussianFunctionTest.patch Currently, the GaussianFitter class contains a CurveFitter instance (and delegates most method calls to it) for the sake of being able to return a GaussianFunction instance (from the call to the fit method). I think that it would clearer (in the context of the fitting package) that fit returns an array of doubles (as CurveFitter does). The user can then easily create a Gaussian function object, if needed. [Also, note that the GaussianFitter actually fits the sum of a constant and a Gaussian function; thus the name is slightly misleading IMO.] I also think that ParametricGaussianFunction and GaussianParametersGuesser should become inner classes of GaussianFitter (as this makes it clear what their use is). -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Resolved: (CONFIGURATION-363) Consider enhancing to allow nested variable interpolation
[
https://issues.apache.org/jira/browse/CONFIGURATION-363?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Oliver Heger resolved CONFIGURATION-363.
Resolution: Fixed
The nested interpolation feature is available with Commons Lang 2.6. A unit
test was added to verify that it actually works.
However, in the {{StrSubstitutor}} class of Commons Lang the feature is
disabled by default. In order to stay compatible with older Lang versions, I
did not want to use an API which is available in version 2.6 only.
So the configuration API does not provide a convenience method to enable nested
interpolation. Rather, you have to switch on this mode manually on the
substitutor used by the configuration. This can be done as follows:
{code}
AbstractConfiguration config = ...
config.getSubstitutor().setEnableSubstitutionInVariables(true);
{code}
Consider enhancing to allow nested variable interpolation
-
Key: CONFIGURATION-363
URL: https://issues.apache.org/jira/browse/CONFIGURATION-363
Project: Commons Configuration
Issue Type: Improvement
Components: Interpolation
Reporter: Rob Walker
Assignee: Oliver Heger
Priority: Minor
Fix For: 1.7
It would be really handy of Commons Configuration could support nested
variable interpolation:
org.osgi.framework.system.packages=org.osgi.framework; version=1.4.0, \
org.osgi.service.packageadmin; version=1.2.0, \
org.osgi.service.startlevel; version=1.1.0, \
org.osgi.service.url; version=1.0.0, \
org.osgi.util.tracker; version=1.3.3 \
${jre-${java.specification.version}}
The process being to expand innermost variable references first and work
outwards - this allows for very advance config, such as the above, which can
dynamically detect the Java version and expand a JRE version specific
property into the property being defined.
Care is needed to avoid recursion, which can been handled through use of a
Cycle Map that detects of an inner variable has already been referenced.
An example of this can be found in the substVars method of the Apache Felix
org.apache.felix.framework.util.Util class.
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Resolved: (MATH-514) Replace the ParametricGaussianFunction class
[ https://issues.apache.org/jira/browse/MATH-514?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Gilles resolved MATH-514. - Resolution: Fixed Class removed in revision 1072207. Replace the ParametricGaussianFunction class -- Key: MATH-514 URL: https://issues.apache.org/jira/browse/MATH-514 Project: Commons Math Issue Type: Improvement Reporter: Gilles Assignee: Gilles Priority: Minor Labels: api-change Fix For: 3.0 The ParametricGaussianFunction is defined in package optimization.fitting whereas it is not specific to the fitting concept. It would be more logical to have that functionality provided as an inner class of the Gaussian class (in package analysis.function). Moreover, no unit test in ParametricGaussianFunctionTest actually tests the code of the value and gradient methods, while tests in that class duplicate some of the tests in GaussianFitterTest (thus testing the fitting code instead of the function). -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Created: (MATH-516) Remove GaussianFunction and GaussianDerivativeFunction
Remove GaussianFunction and GaussianDerivativeFunction -- Key: MATH-516 URL: https://issues.apache.org/jira/browse/MATH-516 Project: Commons Math Issue Type: Improvement Reporter: Gilles Assignee: Gilles Priority: Trivial Fix For: 3.0 Those two classes are not used anymore as their functionality is now contained in class Gaussian (in package analysis.function). -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Resolved: (MATH-516) Remove GaussianFunction and GaussianDerivativeFunction
[ https://issues.apache.org/jira/browse/MATH-516?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Gilles resolved MATH-516. - Resolution: Fixed Revision 1072209. Remove GaussianFunction and GaussianDerivativeFunction -- Key: MATH-516 URL: https://issues.apache.org/jira/browse/MATH-516 Project: Commons Math Issue Type: Improvement Reporter: Gilles Assignee: Gilles Priority: Trivial Labels: api-change Fix For: 3.0 Those two classes are not used anymore as their functionality is now contained in class Gaussian (in package analysis.function). -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Created: (MATH-517) Refactor HarmonicFitter
Refactor HarmonicFitter - Key: MATH-517 URL: https://issues.apache.org/jira/browse/MATH-517 Project: Commons Math Issue Type: Improvement Reporter: Gilles Priority: Minor Fix For: 3.0 HarmonicFitter should be refactored in the same way as GaussianFitter (MATH-512). -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Created: (MATH-518) Move HarmonicFunction to package analysis.function
Move HarmonicFunction to package analysis.function -- Key: MATH-518 URL: https://issues.apache.org/jira/browse/MATH-518 Project: Commons Math Issue Type: Improvement Reporter: Gilles Assignee: Gilles Priority: Minor Fix For: 3.0 The class HarmonicFunction (in package optimization.fitting) should be renamed HarmonicOscillator and moved to package analysis.function. -- This message is automatically generated by JIRA. - For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Created: (MATH-519) GaussianFitter Unexpectedly Throws NotStrictlyPositiveException
GaussianFitter Unexpectedly Throws NotStrictlyPositiveException
---
Key: MATH-519
URL: https://issues.apache.org/jira/browse/MATH-519
Project: Commons Math
Issue Type: Bug
Affects Versions: 3.0
Reporter: Ole Ersoy
Fix For: 3.0
Running the following:
double[] observations =
{
1.1143831578403364E-29,
4.95281403484594E-28,
1.1171347211930288E-26,
1.7044813962636277E-25,
1.9784716574832164E-24,
1.8630236407866774E-23,
1.4820532905097742E-22,
1.0241963854632831E-21,
6.275077366673128E-21,
3.461808994532493E-20,
1.7407124684715706E-19,
8.056687953553974E-19,
3.460193945992071E-18,
1.3883326374011525E-17,
5.233894983671116E-17,
1.8630791465263745E-16,
6.288759227922111E-16,
2.0204433920597856E-15,
6.198768938576155E-15,
1.821419346860626E-14,
5.139176445538471E-14,
1.3956427429045787E-13,
3.655705706448139E-13,
9.253753324779779E-13,
2.267636001476696E-12,
5.3880460095836855E-12,
1.2431632654852931E-11
};
GaussianFitter g =
new GaussianFitter(new LevenbergMarquardtOptimizer());
for (int index = 0; index 27; index++)
{
g.addObservedPoint(index, observations[index]);
}
g.fit();
Results in:
org.apache.commons.math.exception.NotStrictlyPositiveException: -1.277 is
smaller than, or equal to, the minimum (0)
at
org.apache.commons.math.analysis.function.Gaussian$Parametric.validateParameters(Gaussian.java:184)
at
org.apache.commons.math.analysis.function.Gaussian$Parametric.value(Gaussian.java:129)
I'm guessing the initial guess for sigma is off.
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
