Git Mirror of Weka

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

Weka (mirror)

NOTE The owner of this repository has no affiliation with official WEKA project. This repo is periodically updated as a kindness to others who have shown interest in it. It can take several hours to checkout the full official WEKA subversion repository and several minutes just to update with any new commits. Therefore, this exists to provide an easy way to access and peruse the WEKA source using git, nothing more.

---

This is a git mirror of the The University of Waikato machine learning project WEKA.

Weka is a collection of machine learning algorithms for data mining tasks. The algorithms can either be applied directly to a dataset or called from your own Java code. Weka contains tools for data pre-processing, classification, regression, clustering, association rules, and visualization. It is also well-suited for developing new machine learning schemes.

The official WEKA source code is hosted using subversion at the Waikato SVN server.

The current version of WEKA is licensed under the GNU General Public license version 3.0.

Use the weka-trunk branch to follow the official code base under active development at the University of Waikato.

本帖隐藏的内容

https://github.com/bnjmn/weka

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏0 回帖

关键词：Mirror MIRR WEKA IRR periodically nothing commits University checkout learning

1. /*
   
2. *    This program is free software; you can redistribute it and/or modify
   
3. *    it under the terms of the GNU General Public License as published by
   
4. *    the Free Software Foundation; either version 2 of the License, or
   
5. *    (at your option) any later version.
   
6. *
   
7. *    This program is distributed in the hope that it will be useful,
   
8. *    but WITHOUT ANY WARRANTY; without even the implied warranty of
   
9. *    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   
10. *    GNU General Public License for more details.
    
11. *
    
12. *    You should have received a copy of the GNU General Public License
    
13. *    along with this program; if not, write to the Free Software
    
14. *    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
    
15. */
    
16. 
    
17. /*
    
18. *    AttributeSelectionTest.java
    
19. *    Copyright (C) 2009 University of Waikato, Hamilton, New Zealand
    
20. *
    
21. */
    
22. 
    
23. package wekaexamples.attributeSelection;
    
24. 
    
25. import weka.attributeSelection.AttributeSelection;
    
26. import weka.attributeSelection.CfsSubsetEval;
    
27. import weka.attributeSelection.GreedyStepwise;
    
28. import weka.classifiers.Evaluation;
    
29. import weka.classifiers.meta.AttributeSelectedClassifier;
    
30. import weka.classifiers.trees.J48;
    
31. import weka.core.Instances;
    
32. import weka.core.Utils;
    
33. import weka.core.converters.ConverterUtils.DataSource;
    
34. import weka.filters.Filter;
    
35. 
    
36. import java.util.Random;
    
37. 
    
38. /**
    
39. * performs attribute selection using CfsSubsetEval and GreedyStepwise
    
40. * (backwards) and trains J48 with that. Needs 3.5.5 or higher to compile.
    
41. *
    
42. * @author FracPete (fracpete at waikato dot ac dot nz)
    
43. * @version $Revision$
    
44. */
    
45. public class AttributeSelectionTest {
    
46. 
    
47.   /**
    
48.    * uses the meta-classifier
    
49.    */
    
50.   protected static void useClassifier(Instances data) throws Exception {
    
51.     System.out.println("\n1. Meta-classfier");
    
52.     AttributeSelectedClassifier classifier = new AttributeSelectedClassifier();
    
53.     CfsSubsetEval eval = new CfsSubsetEval();
    
54.     GreedyStepwise search = new GreedyStepwise();
    
55.     search.setSearchBackwards(true);
    
56.     J48 base = new J48();
    
57.     classifier.setClassifier(base);
    
58.     classifier.setEvaluator(eval);
    
59.     classifier.setSearch(search);
    
60.     Evaluation evaluation = new Evaluation(data);
    
61.     evaluation.crossValidateModel(classifier, data, 10, new Random(1));
    
62.     System.out.println(evaluation.toSummaryString());
    
63.   }
    
64. 
    
65.   /**
    
66.    * uses the filter
    
67.    */
    
68.   protected static void useFilter(Instances data) throws Exception {
    
69.     System.out.println("\n2. Filter");
    
70.     weka.filters.supervised.attribute.AttributeSelection filter = new weka.filters.supervised.attribute.AttributeSelection();
    
71.     CfsSubsetEval eval = new CfsSubsetEval();
    
72.     GreedyStepwise search = new GreedyStepwise();
    
73.     search.setSearchBackwards(true);
    
74.     filter.setEvaluator(eval);
    
75.     filter.setSearch(search);
    
76.     filter.setInputFormat(data);
    
77.     Instances newData = Filter.useFilter(data, filter);
    
78.     System.out.println(newData);
    
79.   }
    
80. 
    
81.   /**
    
82.    * uses the low level approach
    
83.    */
    
84.   protected static void useLowLevel(Instances data) throws Exception {
    
85.     System.out.println("\n3. Low-level");
    
86.     AttributeSelection attsel = new AttributeSelection();
    
87.     CfsSubsetEval eval = new CfsSubsetEval();
    
88.     GreedyStepwise search = new GreedyStepwise();
    
89.     search.setSearchBackwards(true);
    
90.     attsel.setEvaluator(eval);
    
91.     attsel.setSearch(search);
    
92.     attsel.SelectAttributes(data);
    
93.     int[] indices = attsel.selectedAttributes();
    
94.     System.out.println("selected attribute indices (starting with 0):\n" + Utils.arrayToString(indices));
    
95.   }
    
96. 
    
97.   /**
    
98.    * takes a dataset as first argument
    
99.    *
    
100.    * @param args        the commandline arguments
     
101.    * @throws Exception  if something goes wrong
     
102.    */
     
103.   public static void main(String[] args) throws Exception {
     
104.     // load data
     
105.     System.out.println("\n0. Loading data");
     
106.     DataSource source = new DataSource(args[0]);
     
107.     Instances data = source.getDataSet();
     
108.     if (data.classIndex() == -1)
     
109.       data.setClassIndex(data.numAttributes() - 1);
     
110. 
     
111.     // 1. meta-classifier
     
112.     useClassifier(data);
     
113. 
     
114.     // 2. filter
     
115.     useFilter(data);
     
116. 
     
117.     // 3. low-level
     
118.     useLowLevel(data);
     
119.   }
     
120. }

复制代码

1. /*
   
2. *    This program is free software; you can redistribute it and/or modify
   
3. *    it under the terms of the GNU General Public License as published by
   
4. *    the Free Software Foundation; either version 2 of the License, or
   
5. *    (at your option) any later version.
   
6. *
   
7. *    This program is distributed in the hope that it will be useful,
   
8. *    but WITHOUT ANY WARRANTY; without even the implied warranty of
   
9. *    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   
10. *    GNU General Public License for more details.
    
11. *
    
12. *    You should have received a copy of the GNU General Public License
    
13. *    along with this program; if not, write to the Free Software
    
14. *    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
    
15. */
    
16. 
    
17. /*
    
18. *    AprioriOutput.java
    
19. *    Copyright (C) 2009 University of Waikato, Hamilton, New Zealand
    
20. *
    
21. */
    
22. 
    
23. package wekaexamples.associations;
    
24. 
    
25. import weka.associations.Apriori;
    
26. import weka.core.Instances;
    
27. import weka.core.converters.ConverterUtils.DataSource;
    
28. 
    
29. /**
    
30. * Loads a dataset, builds Apriori on it and outputs Apriori's model.
    
31. *
    
32. * @author FracPete (fracpete at waikato dot ac dot nz)
    
33. * @version $Revision$
    
34. */
    
35. public class AprioriOutput {
    
36. 
    
37.   /**
    
38.    * Expects a dataset as first parameter. The last attribute is used
    
39.    * as class attribute.
    
40.    *
    
41.    * @param args        the command-line parameters
    
42.    * @throws Exception        if something goes wrong
    
43.    */
    
44.   public static void main(String[] args) throws Exception {
    
45.     // load data
    
46.     Instances data = DataSource.read(args[0]);
    
47.     data.setClassIndex(data.numAttributes() - 1);
    
48. 
    
49.     // build associator
    
50.     Apriori apriori = new Apriori();
    
51.     apriori.setClassIndex(data.classIndex());
    
52.     apriori.buildAssociations(data);
    
53. 
    
54.     // output associator
    
55.     System.out.println(apriori);
    
56.   }
    
57. }

复制代码

1. /*
   
2. *    IncrementalClassifier.java
   
3. *    Copyright (C) 2009 University of Waikato, Hamilton, New Zealand
   
4. *
   
5. */
   
6. 
   
7. package wekaexamples.classifiers;
   
8. 
   
9. import weka.classifiers.bayes.NaiveBayesUpdateable;
   
10. import weka.core.Instance;
    
11. import weka.core.Instances;
    
12. import weka.core.converters.ArffLoader;
    
13. 
    
14. import java.io.File;
    
15. 
    
16. /**
    
17. * This example trains NaiveBayes incrementally on data obtained
    
18. * from the ArffLoader.
    
19. *
    
20. * @author FracPete (fracpete at waikato dot ac dot nz)
    
21. * @version $Revision$
    
22. */
    
23. public class IncrementalClassifier {
    
24. 
    
25.   /**
    
26.    * Expects an ARFF file as first argument (class attribute is assumed
    
27.    * to be the last attribute).
    
28.    *
    
29.    * @param args        the commandline arguments
    
30.    * @throws Exception  if something goes wrong
    
31.    */
    
32.   public static void main(String[] args) throws Exception {
    
33.     // load data
    
34.     ArffLoader loader = new ArffLoader();
    
35.     loader.setFile(new File(args[0]));
    
36.     Instances structure = loader.getStructure();
    
37.     structure.setClassIndex(structure.numAttributes() - 1);
    
38. 
    
39.     // train NaiveBayes
    
40.     NaiveBayesUpdateable nb = new NaiveBayesUpdateable();
    
41.     nb.buildClassifier(structure);
    
42.     Instance current;
    
43.     while ((current = loader.getNextInstance(structure)) != null)
    
44.       nb.updateClassifier(current);
    
45. 
    
46.     // output generated model
    
47.     System.out.println(nb);
    
48.   }
    
49. }

复制代码

1. /*
   
2. *    This program is free software; you can redistribute it and/or modify
   
3. *    it under the terms of the GNU General Public License as published by
   
4. *    the Free Software Foundation; either version 2 of the License, or
   
5. *    (at your option) any later version.
   
6. *
   
7. *    This program is distributed in the hope that it will be useful,
   
8. *    but WITHOUT ANY WARRANTY; without even the implied warranty of
   
9. *    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   
10. *    GNU General Public License for more details.
    
11. *
    
12. *    You should have received a copy of the GNU General Public License
    
13. *    along with this program; if not, write to the Free Software
    
14. *    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
    
15. */
    
16. 
    
17. /*
    
18. *    CrossValidationAddPrediction.java
    
19. *    Copyright (C) 2009 University of Waikato, Hamilton, New Zealand
    
20. *
    
21. */
    
22. 
    
23. package wekaexamples.classifiers;
    
24. 
    
25. import weka.classifiers.AbstractClassifier;
    
26. import weka.classifiers.Classifier;
    
27. import weka.classifiers.Evaluation;
    
28. import weka.core.Instances;
    
29. import weka.core.OptionHandler;
    
30. import weka.core.Utils;
    
31. import weka.core.converters.ConverterUtils.DataSink;
    
32. import weka.core.converters.ConverterUtils.DataSource;
    
33. import weka.filters.Filter;
    
34. import weka.filters.supervised.attribute.AddClassification;
    
35. 
    
36. import java.util.Random;
    
37. 
    
38. /**
    
39. * Performs a single run of cross-validation and adds the prediction on the
    
40. * test set to the dataset.
    
41. *
    
42. * Command-line parameters:
    
43. * <ul>
    
44. *    <li>-t filename - the dataset to use</li>
    
45. *    <li>-o filename - the output file to store dataset with the predictions
    
46. *    in</li>
    
47. *    <li>-x int - the number of folds to use</li>
    
48. *    <li>-s int - the seed for the random number generator</li>
    
49. *    <li>-c int - the class index, "first" and "last" are accepted as well;
    
50. *    "last" is used by default</li>
    
51. *    <li>-W classifier - classname and options, enclosed by double quotes;
    
52. *    the classifier to cross-validate</li>
    
53. * </ul>
    
54. *
    
55. * Example command-line:
    
56. * <pre>
    
57. * java wekaexamples.classifiers.CrossValidationAddPrediction -t anneal.arff -c last -o predictions.arff -x 10 -s 1 -W "weka.classifiers.trees.J48 -C 0.25"
    
58. * </pre>
    
59. *
    
60. * @author FracPete (fracpete at waikato dot ac dot nz)
    
61. * @version $Revision$
    
62. */
    
63. public class CrossValidationAddPrediction {
    
64. 
    
65.   /**
    
66.    * Performs the cross-validation. See Javadoc of class for information
    
67.    * on command-line parameters.
    
68.    *
    
69.    * @param args        the command-line parameters
    
70.    * @throws Exception        if something goes wrong
    
71.    */
    
72.   public static void main(String[] args) throws Exception {
    
73.     // loads data and set class index
    
74.     Instances data = DataSource.read(Utils.getOption("t", args));
    
75.     String clsIndex = Utils.getOption("c", args);
    
76.     if (clsIndex.length() == 0)
    
77.       clsIndex = "last";
    
78.     if (clsIndex.equals("first"))
    
79.       data.setClassIndex(0);
    
80.     else if (clsIndex.equals("last"))
    
81.       data.setClassIndex(data.numAttributes() - 1);
    
82.     else
    
83.       data.setClassIndex(Integer.parseInt(clsIndex) - 1);
    
84. 
    
85.     // classifier
    
86.     String[] tmpOptions;
    
87.     String classname;
    
88.     tmpOptions     = Utils.splitOptions(Utils.getOption("W", args));
    
89.     classname      = tmpOptions[0];
    
90.     tmpOptions[0]  = "";
    
91.     Classifier cls = (Classifier) Utils.forName(Classifier.class, classname, tmpOptions);
    
92. 
    
93.     // other options
    
94.     int seed  = Integer.parseInt(Utils.getOption("s", args));
    
95.     int folds = Integer.parseInt(Utils.getOption("x", args));
    
96. 
    
97.     // randomize data
    
98.     Random rand = new Random(seed);
    
99.     Instances randData = new Instances(data);
    
100.     randData.randomize(rand);
     
101.     if (randData.classAttribute().isNominal())
     
102.       randData.stratify(folds);
     
103. 
     
104.     // perform cross-validation and add predictions
     
105.     Instances predictedData = null;
     
106.     Evaluation eval = new Evaluation(randData);
     
107.     for (int n = 0; n < folds; n++) {
     
108.       Instances train = randData.trainCV(folds, n);
     
109.       Instances test = randData.testCV(folds, n);
     
110.       // the above code is used by the StratifiedRemoveFolds filter, the
     
111.       // code below by the Explorer/Experimenter:
     
112.       // Instances train = randData.trainCV(folds, n, rand);
     
113. 
     
114.       // build and evaluate classifier
     
115.       Classifier clsCopy = AbstractClassifier.makeCopy(cls);
     
116.       clsCopy.buildClassifier(train);
     
117.       eval.evaluateModel(clsCopy, test);
     
118. 
     
119.       // add predictions
     
120.       AddClassification filter = new AddClassification();
     
121.       filter.setClassifier(cls);
     
122.       filter.setOutputClassification(true);
     
123.       filter.setOutputDistribution(true);
     
124.       filter.setOutputErrorFlag(true);
     
125.       filter.setInputFormat(train);
     
126.       Filter.useFilter(train, filter);  // trains the classifier
     
127.       Instances pred = Filter.useFilter(test, filter);  // perform predictions on test set
     
128.       if (predictedData == null)
     
129.         predictedData = new Instances(pred, 0);
     
130.       for (int j = 0; j < pred.numInstances(); j++)
     
131.         predictedData.add(pred.instance(j));
     
132.     }
     
133. 
     
134.     // output evaluation
     
135.     System.out.println();
     
136.     System.out.println("=== Setup ===");
     
137.     if (cls instanceof OptionHandler)
     
138.       System.out.println("Classifier: " + cls.getClass().getName() + " " + Utils.joinOptions(((OptionHandler) cls).getOptions()));
     
139.     else
     
140.       System.out.println("Classifier: " + cls.getClass().getName());
     
141.     System.out.println("Dataset: " + data.relationName());
     
142.     System.out.println("Folds: " + folds);
     
143.     System.out.println("Seed: " + seed);
     
144.     System.out.println();
     
145.     System.out.println(eval.toSummaryString("=== " + folds + "-fold Cross-validation ===", false));
     
146. 
     
147.     // output "enriched" dataset
     
148.     DataSink.write(Utils.getOption("o", args), predictedData);
     
149.   }
     
150. }

复制代码

1. /*
   
2. *    This program is free software; you can redistribute it and/or modify
   
3. *    it under the terms of the GNU General Public License as published by
   
4. *    the Free Software Foundation; either version 2 of the License, or
   
5. *    (at your option) any later version.
   
6. *
   
7. *    This program is distributed in the hope that it will be useful,
   
8. *    but WITHOUT ANY WARRANTY; without even the implied warranty of
   
9. *    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   
10. *    GNU General Public License for more details.
    
11. *
    
12. *    You should have received a copy of the GNU General Public License
    
13. *    along with this program; if not, write to the Free Software
    
14. *    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
    
15. */
    
16. 
    
17. /*
    
18. *    CrossValidationAddPrediction.java
    
19. *    Copyright (C) 2009 University of Waikato, Hamilton, New Zealand
    
20. *
    
21. */
    
22. 
    
23. package wekaexamples.classifiers;
    
24. 
    
25. import weka.classifiers.AbstractClassifier;
    
26. import weka.classifiers.Classifier;
    
27. import weka.classifiers.Evaluation;
    
28. import weka.core.Instances;
    
29. import weka.core.OptionHandler;
    
30. import weka.core.Utils;
    
31. import weka.core.converters.ConverterUtils.DataSink;
    
32. import weka.core.converters.ConverterUtils.DataSource;
    
33. import weka.filters.Filter;
    
34. import weka.filters.supervised.attribute.AddClassification;
    
35. 
    
36. import java.util.Random;
    
37. 
    
38. /**
    
39. * Performs a single run of cross-validation and adds the prediction on the
    
40. * test set to the dataset.
    
41. *
    
42. * Command-line parameters:
    
43. * <ul>
    
44. *    <li>-t filename - the dataset to use</li>
    
45. *    <li>-o filename - the output file to store dataset with the predictions
    
46. *    in</li>
    
47. *    <li>-x int - the number of folds to use</li>
    
48. *    <li>-s int - the seed for the random number generator</li>
    
49. *    <li>-c int - the class index, "first" and "last" are accepted as well;
    
50. *    "last" is used by default</li>
    
51. *    <li>-W classifier - classname and options, enclosed by double quotes;
    
52. *    the classifier to cross-validate</li>
    
53. * </ul>
    
54. *
    
55. * Example command-line:
    
56. * <pre>
    
57. * java wekaexamples.classifiers.CrossValidationAddPrediction -t anneal.arff -c last -o predictions.arff -x 10 -s 1 -W "weka.classifiers.trees.J48 -C 0.25"
    
58. * </pre>
    
59. *
    
60. * @author FracPete (fracpete at waikato dot ac dot nz)
    
61. * @version $Revision$
    
62. */
    
63. public class CrossValidationAddPrediction {
    
64. 
    
65.   /**
    
66.    * Performs the cross-validation. See Javadoc of class for information
    
67.    * on command-line parameters.
    
68.    *
    
69.    * @param args        the command-line parameters
    
70.    * @throws Exception        if something goes wrong
    
71.    */
    
72.   public static void main(String[] args) throws Exception {
    
73.     // loads data and set class index
    
74.     Instances data = DataSource.read(Utils.getOption("t", args));
    
75.     String clsIndex = Utils.getOption("c", args);
    
76.     if (clsIndex.length() == 0)
    
77.       clsIndex = "last";
    
78.     if (clsIndex.equals("first"))
    
79.       data.setClassIndex(0);
    
80.     else if (clsIndex.equals("last"))
    
81.       data.setClassIndex(data.numAttributes() - 1);
    
82.     else
    
83.       data.setClassIndex(Integer.parseInt(clsIndex) - 1);
    
84. 
    
85.     // classifier
    
86.     String[] tmpOptions;
    
87.     String classname;
    
88.     tmpOptions     = Utils.splitOptions(Utils.getOption("W", args));
    
89.     classname      = tmpOptions[0];
    
90.     tmpOptions[0]  = "";
    
91.     Classifier cls = (Classifier) Utils.forName(Classifier.class, classname, tmpOptions);
    
92. 
    
93.     // other options
    
94.     int seed  = Integer.parseInt(Utils.getOption("s", args));
    
95.     int folds = Integer.parseInt(Utils.getOption("x", args));
    
96. 
    
97.     // randomize data
    
98.     Random rand = new Random(seed);
    
99.     Instances randData = new Instances(data);
    
100.     randData.randomize(rand);
     
101.     if (randData.classAttribute().isNominal())
     
102.       randData.stratify(folds);
     
103. 
     
104.     // perform cross-validation and add predictions
     
105.     Instances predictedData = null;
     
106.     Evaluation eval = new Evaluation(randData);
     
107.     for (int n = 0; n < folds; n++) {
     
108.       Instances train = randData.trainCV(folds, n);
     
109.       Instances test = randData.testCV(folds, n);
     
110.       // the above code is used by the StratifiedRemoveFolds filter, the
     
111.       // code below by the Explorer/Experimenter:
     
112.       // Instances train = randData.trainCV(folds, n, rand);
     
113. 
     
114.       // build and evaluate classifier
     
115.       Classifier clsCopy = AbstractClassifier.makeCopy(cls);
     
116.       clsCopy.buildClassifier(train);
     
117.       eval.evaluateModel(clsCopy, test);
     
118. 
     
119.       // add predictions
     
120.       AddClassification filter = new AddClassification();
     
121.       filter.setClassifier(cls);
     
122.       filter.setOutputClassification(true);
     
123.       filter.setOutputDistribution(true);
     
124.       filter.setOutputErrorFlag(true);
     
125.       filter.setInputFormat(train);
     
126.       Filter.useFilter(train, filter);  // trains the classifier
     
127.       Instances pred = Filter.useFilter(test, filter);  // perform predictions on test set
     
128.       if (predictedData == null)
     
129.         predictedData = new Instances(pred, 0);
     
130.       for (int j = 0; j < pred.numInstances(); j++)
     
131.         predictedData.add(pred.instance(j));
     
132.     }
     
133. 
     
134.     // output evaluation
     
135.     System.out.println();
     
136.     System.out.println("=== Setup ===");
     
137.     if (cls instanceof OptionHandler)
     
138.       System.out.println("Classifier: " + cls.getClass().getName() + " " + Utils.joinOptions(((OptionHandler) cls).getOptions()));
     
139.     else
     
140.       System.out.println("Classifier: " + cls.getClass().getName());
     
141.     System.out.println("Dataset: " + data.relationName());
     
142.     System.out.println("Folds: " + folds);
     
143.     System.out.println("Seed: " + seed);
     
144.     System.out.println();
     
145.     System.out.println(eval.toSummaryString("=== " + folds + "-fold Cross-validation ===", false));
     
146. 
     
147.     // output "enriched" dataset
     
148.     DataSink.write(Utils.getOption("o", args), predictedData);
     
149.   }
     
150. }

复制代码