sdvillal/fast-random-forest

Good old Fran Supek FRF for weka. Automatically exported from code.google.com/p/fast-random-forest.

fast-random-forest's Issues

Potential Serialization Issue

What steps will reproduce the problem?
1. Use SerialiationHelper to serialize a trained FastRandomForest Classifier 
object

What is the expected output? What do you see instead?

While it does serialize the object, it seems to also be serializing the 
instance data that was used to train the object. This makes the serialized 
files potentially very large (based on training set) and makes loading 
serialized objects take much longer due to having to load a training set when 
it isn't needed.

What version of the product are you using? On what operating system?

0.98 frf
Java 1.6.0_31
Weka 3.6.6
Mac OSX 10.7


Please provide any additional information below.

Most likely storing instance data in some variable that also gets serialized 
later.

Original issue reported on code.google.com by [email protected] on 6 May 2012 at 8:47

Exception in when training set numlasses != test set numClasses

Hope I'm not missing something basic.

What steps will reproduce the problem?
1. //Create a test set with a different number of classes as a training set
2. cls.buildClassifier(trainData);
3. Evaluation eval = new Evaluation(trainData);
4. eval.evaluateModel(cls, testData);

What is the expected output? What do you see instead?

Expected to complete evaluation successfully uith no output.

Instead, step 4 gives exception:

Exception in thread "main" java.lang.ArrayIndexOutOfBoundsException: 13
    hr.irb.fastRandomForest.FastRfBagging.distributionForInstance(FastRfBagging.java:647)
    hr.irb.fastRandomForest.FastRandomForest.distributionForInstance(FastRandomForest.java:644)
    weka.classifiers.evaluation.Evaluation.evaluationForSingleInstance(Evaluation.java:1937)
    weka.classifiers.evaluation.Evaluation.evaluateModelOnceAndRecordPrediction(Evaluation.java:1976)
    weka.classifiers.evaluation.Evaluation.evaluateModel(Evaluation.java:1854)
    weka.classifiers.Evaluation.evaluateModel(Evaluation.java:671)
    cf.PortClassifier.main(PortClassifier.java:85)

    at hr.irb.fastRandomForest.FastRfBagging.distributionForInstance(FastRfBagging.java:647)
    at hr.irb.fastRandomForest.FastRandomForest.distributionForInstance(FastRandomForest.java:644)
    at weka.classifiers.evaluation.Evaluation.evaluationForSingleInstance(Evaluation.java:1937)
    at weka.classifiers.evaluation.Evaluation.evaluateModelOnceAndRecordPrediction(Evaluation.java:1976)
    at weka.classifiers.evaluation.Evaluation.evaluateModel(Evaluation.java:1854)
    at weka.classifiers.Evaluation.evaluateModel(Evaluation.java:671)
    at cf.PortClassifier.main(PortClassifier.java:85)



What version of the product are you using? On what operating system?
0.99 on Ubuntu 14.04

Please provide any additional information below.

Original issue reported on code.google.com by [email protected] on 16 Jan 2015 at 4:16

Update copyright notices to new Free Software Foundation mailing address

The Free Software Foundation is no longer at 675 Mass Ave, Cambridge, MA.

Our new address is 51 Franklin Street, Suite 500, Boston, MA 02110.

You can confirm this for yourself here: http://www.fsf.org/about/contact/

Please update all references to our old address in your code so people can 
continue to contact us (we haven't been at the old address for more than a 
decade at this point)

Kind regards,

matt

--
Matt Lee
Campaigns Manager
Free Software Foundation

Original issue reported on code.google.com by [email protected] on 8 Sep 2011 at 3:33

WEKA 3.7.3

I have compared the fast random forest implementation to the new WEKA 
RandomForest implementation.

Since version 3.7.1 WEKA supports parallel processing for ensemble classifiers. 
I found that the speed-up of the fast-random-forest implementation is marginal, 
while the quality in terms of accuracy on my test data is better using the WEKA 
RandomForest implementation.

Thus, I recommend to use WEKA 3.7.1 instead of this implementation.

Original issue reported on code.google.com by [email protected] on 20 Jan 2011 at 2:22

NPE when using on basic text classifciation

What steps will reproduce the problem?
I'm trying to do a "Hello World" of text classification.  It works fine using 
NaiveBayes.  When I switch to FRF, I get:

Exception in thread "main" java.util.concurrent.ExecutionException: 
java.lang.NullPointerException
    at java.util.concurrent.FutureTask$Sync.innerGet(FutureTask.java:252)
    at java.util.concurrent.FutureTask.get(FutureTask.java:111)
    at hr.irb.fastRandomForest.FastRfBagging.buildClassifier(FastRfBagging.java:172)
    at hr.irb.fastRandomForest.FastRandomForest.buildClassifier(FastRandomForest.java:575)
    at weka.classifiers.meta.FilteredClassifier.buildClassifier(FilteredClassifier.java:442)
    at weka.classifiers.evaluation.Evaluation.crossValidateModel(Evaluation.java:763)
    at weka.classifiers.Evaluation.crossValidateModel(Evaluation.java:373)
    at MyFRF.trainExample(MyFRF.java:87)
    at MyFRF.main(MyFRF.java:48)
Caused by: java.lang.NullPointerException
    at hr.irb.fastRandomForest.FastRandomTree.buildTree(FastRandomTree.java:319)
    at hr.irb.fastRandomForest.FastRandomTree.run(FastRandomTree.java:195)
    at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471)
    at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:334)
    at java.util.concurrent.FutureTask.run(FutureTask.java:166)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
    at java.lang.Thread.run(Thread.java:722)


It happens when I do cross-validation, or plain training.

    classifier.buildClassifier(instances); // HERE

    Evaluation eval = new Evaluation(instances);
    eval.crossValidateModel(classifier, instances, 4, new Random(1)); // OR HERE


What version of the product are you using? On what operating system?
Latest from SVN.  Latest java 1.7

Please provide any additional information below.

Relation Name:  Training Instances
Num Instances:  2890
Num Attributes: 2

     Name                      Type  Nom  Int Real     Missing      Unique  Dist
   1 @@class@@                  Nom 100%   0%   0%     0 /  0%     0 /  0%    40 
   2 text                       Str 100%   0%   0%     0 /  0%  2544 / 88%  2712 

Original issue reported on code.google.com by [email protected] on 14 May 2013 at 5:23

Source code does not compile - AbstractClassifier missing

What steps will reproduce the problem?
Try to compile source code. Add provided weka-3.7.0.jar to build path.

Errors:
AbstractClassifier cannot be resolved to a 
type    Benchmark.java  AbstractClassifier cannot be resolved   FRFAttributeEval.java   
...


Reason:
weka.classifiers.AbstractClassifier
inside weka-3.7.0.jar is missing.


Workaround:

Download and use original weka.jar instead of proivded file.

Original issue reported on code.google.com by [email protected] on 16 Aug 2013 at 6:34

Cannot add it to Weka 3.6.6

According to the instruction in webpage 
'http://code.google.com/p/fast-random-forest/', but I still get a failure to 
add FastRF jar to WEKA 3.6.6.  

Original issue reported on code.google.com by [email protected] on 29 Mar 2012 at 5:08

NullPointerException if classIndex is 0 ( = first Attribute )

What steps will reproduce the problem?
1. Create Instances with first Attribute being the class
2. build a FastRandomForest with these Instances


What version of the product are you using? On what operating system?
0.99 
0.98

Solution
In FastRandomTree replace data.sortedIndices[0].length with 
data.sortedIndices[attIndicesWindow[0]].length

In version 0.99 only one line has to be changed when calling buildTree() in the 
run() method.

In version 0.98 several occurences have to be replaced withing the buildTree 
function.

Original issue reported on code.google.com by [email protected] on 6 May 2014 at 1:18

Does this work with sparse data?

If I'm not mistaken, Weka allows for both dense and sparse dataset files.
Does this work with a sparse dataset, while maintaining a sparse
representation in memory, or must it represent them in memory densely?

Original issue reported on code.google.com by [email protected] on 10 Oct 2009 at 3:40

Requires class attribute to be last

What steps will reproduce the problem?
1. Create an Instances object where the class attribute is not the last 
attribute.
2. Try to build a classifier using FastRandomForest.

What is the expected output? What do you see instead?
NullPointerException

What version of the product are you using? On what operating system?
0.99

Please provide any additional information below.
It works fine if the Instances object is altered to make the class attribute 
last. All classifiers bundled with Weka can handle the class anywhere in the 
list of attributes, so this one should be changed to do that as well.

Original issue reported on code.google.com by [email protected] on 21 Feb 2014 at 2:02

Use from the command-line?

I'm using weka RandomForest from the command-line and tried to substitute
fast-random-forest. 

For example: 

java -cp weka.jar:FastRandomForest.jar
hr.irb.fastRandomforest.FastRandomForest ... lots of options ... -threads
<nthreads> 

and I get:

Exception in thread "main" java.lang.NoClassDefFoundError:
hr/irb/fastRandomforest/FastRandomForest

java -version

java version "1.5.0_15"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0_15-b04)
Java HotSpot(TM) 64-Bit Server VM (build 1.5.0_15-b04, mixed mode)

In case I'm being stupid, can you add a section addressing this use-case?

Thanks,

- n

Original issue reported on code.google.com by [email protected] on 31 May 2009 at 1:44

Build file missing

FastRandomForest cannot be compiled using the source from SVN because
build.xml has been generated by NetBeans and the needed project files for
it cannot be found (ie., there is no nbproject/ directory in the repository).

The build.xml should be either independent of the NetBeans project files or
the NetBeans project files should be added to the repository.

Original issue reported on code.google.com by [email protected] on 18 Feb 2010 at 10:17