keyA.txt

*****NEW RESULT*************************
******KeyA_B.arff***
best classifier: weka.classifiers.functions.MultilayerPerceptron
arguments: [-L, 0.6546470757134388, -M, 0.3401889072466995, -B, -H, o, -C, -R, -S, 1]
attribute search: null
attribute search arguments: []
attribute evaluation: null
attribute evaluation arguments: []
metric: errorRate
estimated errorRate: 0.11914893617021277
training time on evaluation dataset: 0.542 seconds

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

Classifier classifier = AbstractClassifier.forName("weka.classifiers.functions.MultilayerPerceptron", new String[]{"-L", "0.6546470757134388", "-M", "0.3401889072466995", "-B", "-H", "o", "-C", "-R", "-S", "1"});
classifier.buildClassifier(instances);


Correctly Classified Instances         207               88.0851 %
Incorrectly Classified Instances        28               11.9149 %
Kappa statistic                          0.2985
Mean absolute error                      0.1994
Root mean squared error                  0.3204
Relative absolute error                 78.0133 %
Root relative squared error             89.9817 %
Total Number of Instances              235     

=== Confusion Matrix ===

   a   b   <-- classified as
 200   0 |   a = FALSE
  28   7 |   b = TRUE

=== Detailed Accuracy By Class ===

                 TP Rate  FP Rate  Precision  Recall   F-Measure  MCC      ROC Area  PRC Area  Class
                 1.000    0.800    0.877      1.000    0.935      0.419    0.669     0.927     FALSE
                 0.200    0.000    1.000      0.200    0.333      0.419    0.669     0.392     TRUE
Weighted Avg.    0.881    0.681    0.895      0.881    0.845      0.419    0.669     0.847     

Temporary run directories:
/tmp/autoweka6404330954169672090/


For better performance, try giving Auto-WEKA more time.
Tried 558 configurations; to get good results reliably you may need to allow for trying thousands of configurations.

*********************************************
*****NEW RESULT*************************
******KeyA_A.arff***
best classifier: weka.classifiers.meta.RandomSubSpace
arguments: [-I, 36, -P, 0.2757331204049802, -S, 1, -W, weka.classifiers.lazy.IBk, --, -E, -K, 10, -X]
attribute search: null
attribute search arguments: []
attribute evaluation: null
attribute evaluation arguments: []
metric: errorRate
estimated errorRate: 0.16916488222698073
training time on evaluation dataset: 0.208 seconds

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

Classifier classifier = AbstractClassifier.forName("weka.classifiers.meta.RandomSubSpace", new String[]{"-I", "36", "-P", "0.2757331204049802", "-S", "1", "-W", "weka.classifiers.lazy.IBk", "--", "-E", "-K", "10", "-X"});
classifier.buildClassifier(instances);


Correctly Classified Instances         388               83.0835 %
Incorrectly Classified Instances        79               16.9165 %
Kappa statistic                          0.4922
Mean absolute error                      0.2979
Root mean squared error                  0.3505
Relative absolute error                 73.7287 %
Root relative squared error             78.0181 %
Total Number of Instances              467     

=== Confusion Matrix ===

   a   b   <-- classified as
  54  77 |   a = TRUE
   2 334 |   b = FALSE

=== Detailed Accuracy By Class ===

                 TP Rate  FP Rate  Precision  Recall   F-Measure  MCC      ROC Area  PRC Area  Class
                 0.412    0.006    0.964      0.412    0.578      0.562    0.935     0.860     TRUE
                 0.994    0.588    0.813      0.994    0.894      0.562    0.935     0.974     FALSE
Weighted Avg.    0.831    0.425    0.855      0.831    0.805      0.562    0.935     0.942     

Temporary run directories:
/tmp/autoweka8476583998957129575/


For better performance, try giving Auto-WEKA more time.
Tried 431 configurations; to get good results reliably you may need to allow for trying thousands of configurations.

*********************************************
*****NEW RESULT*************************
******KeyA_A_O1.arff***
best classifier: weka.classifiers.trees.RandomForest
arguments: [-I, 111, -K, 1, -depth, 0]
attribute search: weka.attributeSelection.BestFirst
attribute search arguments: [-D, 1, -N, 5]
attribute evaluation: weka.attributeSelection.CfsSubsetEval
attribute evaluation arguments: [-L]
metric: errorRate
estimated errorRate: 0.026785714285714284
training time on evaluation dataset: 0.467 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", "5"});
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.trees.RandomForest", new String[]{"-I", "111", "-K", "1", "-depth", "0"});
classifier.buildClassifier(instances);


Correctly Classified Instances         654               97.3214 %
Incorrectly Classified Instances        18                2.6786 %
Kappa statistic                          0.9464
Mean absolute error                      0.0679
Root mean squared error                  0.1455
Relative absolute error                 13.5766 %
Root relative squared error             29.0938 %
Total Number of Instances              672     

=== Confusion Matrix ===

   a   b   <-- classified as
 330   6 |   a = TRUE
  12 324 |   b = FALSE

=== Detailed Accuracy By Class ===

                 TP Rate  FP Rate  Precision  Recall   F-Measure  MCC      ROC Area  PRC Area  Class
                 0.982    0.036    0.965      0.982    0.973      0.947    0.998     0.998     TRUE
                 0.964    0.018    0.982      0.964    0.973      0.947    0.998     0.998     FALSE
Weighted Avg.    0.973    0.027    0.973      0.973    0.973      0.947    0.998     0.998     

Temporary run directories:
/tmp/autoweka5315849341070509095/


For better performance, try giving Auto-WEKA more time.
Tried 298 configurations; to get good results reliably you may need to allow for trying thousands of configurations.

*********************************************
*****NEW RESULT*************************
******KeyA_A_S1.arff***
best classifier: weka.classifiers.meta.RandomSubSpace
arguments: [-I, 36, -P, 0.2757331204049802, -S, 1, -W, weka.classifiers.lazy.IBk, --, -E, -K, 10, -X]
attribute search: null
attribute search arguments: []
attribute evaluation: null
attribute evaluation arguments: []
metric: errorRate
estimated errorRate: 0.09672619047619048
training time on evaluation dataset: 0.093 seconds

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

Classifier classifier = AbstractClassifier.forName("weka.classifiers.meta.RandomSubSpace", new String[]{"-I", "36", "-P", "0.2757331204049802", "-S", "1", "-W", "weka.classifiers.lazy.IBk", "--", "-E", "-K", "10", "-X"});
classifier.buildClassifier(instances);


Correctly Classified Instances         607               90.3274 %
Incorrectly Classified Instances        65                9.6726 %
Kappa statistic                          0.8065
Mean absolute error                      0.2456
Root mean squared error                  0.2942
Relative absolute error                 49.1244 %
Root relative squared error             58.8316 %
Total Number of Instances              672     

=== Confusion Matrix ===

   a   b   <-- classified as
 287  49 |   a = TRUE
  16 320 |   b = FALSE

=== Detailed Accuracy By Class ===

                 TP Rate  FP Rate  Precision  Recall   F-Measure  MCC      ROC Area  PRC Area  Class
                 0.854    0.048    0.947      0.854    0.898      0.810    0.979     0.981     TRUE
                 0.952    0.146    0.867      0.952    0.908      0.810    0.979     0.979     FALSE
Weighted Avg.    0.903    0.097    0.907      0.903    0.903      0.810    0.979     0.980     

Temporary run directories:
/tmp/autoweka686666325455334675/


For better performance, try giving Auto-WEKA more time.
Tried 410 configurations; to get good results reliably you may need to allow for trying thousands of configurations.

*********************************************
*****NEW RESULT*************************
******KeyA_A_U1.arff***
best classifier: weka.classifiers.meta.Bagging
arguments: [-P, 31, -I, 70, -S, 1, -W, weka.classifiers.trees.DecisionStump, --]
attribute search: weka.attributeSelection.BestFirst
attribute search arguments: [-D, 0, -N, 7]
attribute evaluation: weka.attributeSelection.CfsSubsetEval
attribute evaluation arguments: []
metric: errorRate
estimated errorRate: 0.27099236641221375
training time on evaluation dataset: 0.04 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", "0", "-N", "7"});
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.meta.Bagging", new String[]{"-P", "31", "-I", "70", "-S", "1", "-W", "weka.classifiers.trees.DecisionStump", "--"});
classifier.buildClassifier(instances);


Correctly Classified Instances         191               72.9008 %
Incorrectly Classified Instances        71               27.0992 %
Kappa statistic                          0.458 
Mean absolute error                      0.4451
Root mean squared error                  0.4539
Relative absolute error                 89.0221 %
Root relative squared error             90.7795 %
Total Number of Instances              262     

=== Confusion Matrix ===

   a   b   <-- classified as
 112  19 |   a = TRUE
  52  79 |   b = FALSE

=== Detailed Accuracy By Class ===

                 TP Rate  FP Rate  Precision  Recall   F-Measure  MCC      ROC Area  PRC Area  Class
                 0.855    0.397    0.683      0.855    0.759      0.473    0.801     0.763     TRUE
                 0.603    0.145    0.806      0.603    0.690      0.473    0.801     0.824     FALSE
Weighted Avg.    0.729    0.271    0.745      0.729    0.725      0.473    0.801     0.793     

Temporary run directories:
/tmp/autoweka9208704755207202449/


For better performance, try giving Auto-WEKA more time.
Tried 495 configurations; to get good results reliably you may need to allow for trying thousands of configurations.

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