ant3.txt

*****NEW RESULT*************************
***ant_b.arff***
best classifier: weka.classifiers.rules.DecisionTable
arguments: [-E, acc, -I, -S, weka.attributeSelection.BestFirst, -X, 1]
attribute search: null
attribute search arguments: []
attribute evaluation: null
attribute evaluation arguments: []
metric: errorRate
estimated errorRate: 0.14462081128747795
training time on evaluation dataset: 0.168 seconds

You can use the chosen classifier in your own code as follows:

Classifier classifier = AbstractClassifier.forName("weka.classifiers.rules.DecisionTable", new String[]{"-E", "acc", "-I", "-S", "weka.attributeSelection.BestFirst", "-X", "1"});
classifier.buildClassifier(instances);


Correctly Classified Instances         485               85.5379 %
Incorrectly Classified Instances        82               14.4621 %
Kappa statistic                          0.5653
Mean absolute error                      0.2387
Root mean squared error                  0.3388
Relative absolute error                 66.7283 %
Root relative squared error             80.1747 %
Total Number of Instances              567     

=== Confusion Matrix ===

   a   b   <-- classified as
 407  28 |   a = FALSE
  54  78 |   b = TRUE

=== Detailed Accuracy By Class ===

                 TP Rate  FP Rate  Precision  Recall   F-Measure  MCC      ROC Area  PRC Area  Class
                 0.936    0.409    0.883      0.936    0.908      0.571    0.846     0.926     FALSE
                 0.591    0.064    0.736      0.591    0.655      0.571    0.846     0.630     TRUE
Weighted Avg.    0.855    0.329    0.849      0.855    0.850      0.571    0.846     0.857     

Temporary run directories:
/tmp/autoweka3975115486730593761/


For better performance, try giving Auto-WEKA more time.
Tried 411 configurations; to get good results reliably you may need to allow for trying thousands of configurations.

*********************************************
*****NEW RESULT*************************
***ant_B.arff***
best classifier: weka.classifiers.functions.MultilayerPerceptron
arguments: [-L, 0.2737810261424545, -M, 0.8248143427139586, -B, -H, t, -C, -D, -S, 1]
attribute search: weka.attributeSelection.BestFirst
attribute search arguments: [-D, 1, -N, 9]
attribute evaluation: weka.attributeSelection.CfsSubsetEval
attribute evaluation arguments: []
metric: errorRate
estimated errorRate: 0.17066666666666666
training time on evaluation dataset: 4.682 seconds

You can use the chosen classifier in your own code as follows:

AttributeSelection as = new AttributeSelection();
ASSearch asSearch = ASSearch.forName("weka.attributeSelection.BestFirst", new String[]{"-D", "1", "-N", "9"});
as.setSearch(asSearch);
ASEvaluation asEval = ASEvaluation.forName("weka.attributeSelection.CfsSubsetEval", new String[]{});
as.setEvaluator(asEval);
as.SelectAttributes(instances);
instances = as.reduceDimensionality(instances);
Classifier classifier = AbstractClassifier.forName("weka.classifiers.functions.MultilayerPerceptron", new String[]{"-L", "0.2737810261424545", "-M", "0.8248143427139586", "-B", "-H", "t", "-C", "-D", "-S", "1"});
classifier.buildClassifier(instances);


Correctly Classified Instances         933               82.9333 %
Incorrectly Classified Instances       192               17.0667 %
Kappa statistic                          0.3385
Mean absolute error                      0.2448
Root mean squared error                  0.3491
Relative absolute error                 78.2619 %
Root relative squared error             88.3212 %
Total Number of Instances             1125     

=== Confusion Matrix ===

   a   b   <-- classified as
 861  46 |   a = FALSE
 146  72 |   b = TRUE

=== Detailed Accuracy By Class ===

                 TP Rate  FP Rate  Precision  Recall   F-Measure  MCC      ROC Area  PRC Area  Class
                 0.949    0.670    0.855      0.949    0.900      0.361    0.796     0.925     FALSE
                 0.330    0.051    0.610      0.330    0.429      0.361    0.796     0.545     TRUE
Weighted Avg.    0.829    0.550    0.808      0.829    0.808      0.361    0.796     0.852     

Temporary run directories:
/tmp/autoweka8473494960620495967/


For better performance, try giving Auto-WEKA more time.
Tried 275 configurations; to get good results reliably you may need to allow for trying thousands of configurations.

*********************************************
*****NEW RESULT*************************
***ant_A_O1.arff***
best classifier: weka.classifiers.meta.AdaBoostM1
arguments: [-P, 100, -I, 23, -S, 1, -W, weka.classifiers.rules.PART, --, -M, 6, -B]
attribute search: weka.attributeSelection.GreedyStepwise
attribute search arguments: [-N, 213]
attribute evaluation: weka.attributeSelection.CfsSubsetEval
attribute evaluation arguments: [-L]
metric: errorRate
estimated errorRate: 0.024255788313120176
training time on evaluation dataset: 2.469 seconds

You can use the chosen classifier in your own code as follows:

AttributeSelection as = new AttributeSelection();
ASSearch asSearch = ASSearch.forName("weka.attributeSelection.GreedyStepwise", new String[]{"-N", "213"});
as.setSearch(asSearch);
ASEvaluation asEval = ASEvaluation.forName("weka.attributeSelection.CfsSubsetEval", new String[]{"-L"});
as.setEvaluator(asEval);
as.SelectAttributes(instances);
instances = as.reduceDimensionality(instances);
Classifier classifier = AbstractClassifier.forName("weka.classifiers.meta.AdaBoostM1", new String[]{"-P", "100", "-I", "23", "-S", "1", "-W", "weka.classifiers.rules.PART", "--", "-M", "6", "-B"});
classifier.buildClassifier(instances);


Correctly Classified Instances        1770               97.5744 %
Incorrectly Classified Instances        44                2.4256 %
Kappa statistic                          0.9515
Mean absolute error                      0.0407
Root mean squared error                  0.1234
Relative absolute error                  8.1489 %
Root relative squared error             24.675  %
Total Number of Instances             1814     

=== Confusion Matrix ===

   a   b   <-- classified as
 882  25 |   a = FALSE
  19 888 |   b = TRUE

=== Detailed Accuracy By Class ===

                 TP Rate  FP Rate  Precision  Recall   F-Measure  MCC      ROC Area  PRC Area  Class
                 0.972    0.021    0.979      0.972    0.976      0.952    0.999     0.999     FALSE
                 0.979    0.028    0.973      0.979    0.976      0.952    0.999     0.999     TRUE
Weighted Avg.    0.976    0.024    0.976      0.976    0.976      0.952    0.999     0.999     

Temporary run directories:
/tmp/autoweka8243186695544908750/


For better performance, try giving Auto-WEKA more time.
Tried 189 configurations; to get good results reliably you may need to allow for trying thousands of configurations.

*********************************************
*****NEW RESULT*************************
***ant_A_S1.arff***
best classifier: weka.classifiers.bayes.BayesNet
arguments: [-Q, weka.classifiers.bayes.net.search.local.TAN]
attribute search: null
attribute search arguments: []
attribute evaluation: null
attribute evaluation arguments: []
metric: errorRate
estimated errorRate: 0.11466372657111357
training time on evaluation dataset: 0.13 seconds

You can use the chosen classifier in your own code as follows:

Classifier classifier = AbstractClassifier.forName("weka.classifiers.bayes.BayesNet", new String[]{"-Q", "weka.classifiers.bayes.net.search.local.TAN"});
classifier.buildClassifier(instances);


Correctly Classified Instances        1609               88.699  %
Incorrectly Classified Instances       205               11.301  %
Kappa statistic                          0.774 
Mean absolute error                      0.1554
Root mean squared error                  0.2979
Relative absolute error                 31.0798 %
Root relative squared error             59.5795 %
Total Number of Instances             1814     

=== Confusion Matrix ===

   a   b   <-- classified as
 855  52 |   a = FALSE
 153 754 |   b = TRUE

=== Detailed Accuracy By Class ===

                 TP Rate  FP Rate  Precision  Recall   F-Measure  MCC      ROC Area  PRC Area  Class
                 0.943    0.169    0.848      0.943    0.893      0.779    0.948     0.939     FALSE
                 0.831    0.057    0.935      0.831    0.880      0.779    0.948     0.957     TRUE
Weighted Avg.    0.887    0.113    0.892      0.887    0.887      0.779    0.948     0.948     

Temporary run directories:
/tmp/autoweka7003603607638932432/


For better performance, try giving Auto-WEKA more time.
Tried 184 configurations; to get good results reliably you may need to allow for trying thousands of configurations.

*********************************************
*****NEW RESULT*************************
***ant_A_U1.arff***
best classifier: weka.classifiers.functions.Logistic
arguments: [-R, 1.0469091161449093E-11]
attribute search: weka.attributeSelection.GreedyStepwise
attribute search arguments: [-B, -N, 54]
attribute evaluation: weka.attributeSelection.CfsSubsetEval
attribute evaluation arguments: []
metric: errorRate
estimated errorRate: 0.2545871559633027
training time on evaluation dataset: 0.067 seconds

You can use the chosen classifier in your own code as follows:

AttributeSelection as = new AttributeSelection();
ASSearch asSearch = ASSearch.forName("weka.attributeSelection.GreedyStepwise", new String[]{"-B", "-N", "54"});
as.setSearch(asSearch);
ASEvaluation asEval = ASEvaluation.forName("weka.attributeSelection.CfsSubsetEval", new String[]{});
as.setEvaluator(asEval);
as.SelectAttributes(instances);
instances = as.reduceDimensionality(instances);
Classifier classifier = AbstractClassifier.forName("weka.classifiers.functions.Logistic", new String[]{"-R", "1.0469091161449093E-11"});
classifier.buildClassifier(instances);


Correctly Classified Instances         325               74.5413 %
Incorrectly Classified Instances       111               25.4587 %
Kappa statistic                          0.4908
Mean absolute error                      0.3524
Root mean squared error                  0.4193
Relative absolute error                 70.4839 %
Root relative squared error             83.8509 %
Total Number of Instances              436     

=== Confusion Matrix ===

   a   b   <-- classified as
 171  47 |   a = FALSE
  64 154 |   b = TRUE

=== Detailed Accuracy By Class ===

                 TP Rate  FP Rate  Precision  Recall   F-Measure  MCC      ROC Area  PRC Area  Class
                 0.784    0.294    0.728      0.784    0.755      0.492    0.817     0.804     FALSE
                 0.706    0.216    0.766      0.706    0.735      0.492    0.817     0.810     TRUE
Weighted Avg.    0.745    0.255    0.747      0.745    0.745      0.492    0.817     0.807     

Temporary run directories:
/tmp/autoweka2570685561176935911/


For better performance, try giving Auto-WEKA more time.
Tried 467 configurations; to get good results reliably you may need to allow for trying thousands of configurations.

*********************************************