This note gives an example of how you can use the command-line version of ROSETTA, CLROSETTA, to execute a cross-validation experiment. (The cross-validation can be run from within the ROSETTA GUI, too.) If you are interested in more help resources about CLROSETTA, you might want to examine the Perl script available from the Utilities section of the ROSETTA homepage. In that example, CLROSETTA (and CLHYPOCLASS) are used as computational engines where their input is controlled programmatically by the Perl script. Assume that we want to run a 10-fold cross-validation simulation on the Iris data. (You can find a copy of the Iris data in the Samples folde that accompanies ROSETTA.) For each of the 10 folds, we decide to do the following: In the training stage: 1) Discretize the condition attributes in the training table by using the technique implemented by the BROrthogonalScaler algorithm. 2) If any of the condition attributes were left untouched by the previous step, discretize these using a simple equal frequency binning technique implemented by the EqualFrequencyScaler algorithm. 3) Compute one object-related reduct for each object using the JohnsonReducer algorithm. 4) Generate rules from these reducts using an algorithm in the RuleGenerator family. 5) Using the MyRuleFilter algorithm, filter away all rules that are supported by a single object only. In the test stage: 1) Discretize the condition attributes in the test table by using the cuts computed from the training table in the training stage. This can be done by applying the OrthogonalFileScaler algorithm twice and using the stored cuts. 2) Classify all objects in the discretized test table using the rules computed in the training stage. This can be done by coupling the BatchClassifier algorithm with the StandardVoter algorithm. This can be realized by the following ROSETTA/CLROSETTA script that can be executed by the CVSerialExecutor algorithm: % Training pipeline. BROrthogonalScaler {MASK = False; FILENAME = c:\temp\cuts#ITERATION#.txt} EqualFrequencyScaler {MASK = True; FILENAME = c:\temp\backupcuts#ITERATION#.txt} JohnsonReducer {DISCERNIBILITY = Object; MODULO.DECISION = True; SELECTION = All;} RuleGenerator {} MyRuleFilter {FILTERING = 1; SUPPORT.RHS.LOWER = 0; SUPPORT.RHS.UPPER = 1;} % Testing pipeline OrthogonalFileScaler {MASK = False; FILENAME = c:\temp\cuts#ITERATION#.txt} OrthogonalFileScaler {MASK = True; FILENAME = c:\temp\backupcuts#ITERATION#.txt} BatchClassifier {CLASSIFIER = StandardVoter; VOTING = Support; FALLBACK = True; FALLBACK.CLASS = Iris-setosa; FALLBACK.CERTAINTY = 0.33; MULTIPLE = Best; LOG = True; LOG.FILENAME = c:\temp\log#ITERATION#.txt; LOG.VERBOSE = True; CALIBRATION = False; ROC = False;} We note that the length of the training pipeline is 5, since we will need to give this information to the CVSerialExecutor algorithm as the LENGTH parameter. Assume that we have the decision table with all the raw Iris data in a file called "iris.ros", saved in internal ROSETTA format in the current directory. Furthermore, assume that the script above is saved in "iriscv.txt" in the current directory, and that we want to record a log of our 10-fold cross-validation experiment in the file "iriscv.log" in the current directory. Invoke CLROSETTA as follows from the command-line prompt: clrosetta cvserialexecutor "number = 10; filename.commands = iriscv.txt; length = 5; filename.log = iriscv.log" iris.ros Note that the parameter list is quoted so that it will be treated as a single argument by the shell. After a few seconds the 10-fold CV run should be completed, and various text should have been output to the shell. The contents of the file "iriscv.log" should look something like below. % Output from Rosetta, aleks 1999.04.29 15:43:20 % % CVSerialExecutor % {INVERT=F; NUMBER=10; SEED=0; LENGTH=5; OUTPUT=Undefined; FILENAME.COMMANDS=iriscv.txt; FILENAME.LOG=iriscv.log} 1999.04.29 15:43:20 Executing BROrthogonalScaler... Parameters = {MODE=Save; MASK=F; FILENAME=c:\temp\cuts0.txt; APPROXIMATE=F} 1999.04.29 15:43:21 Output = RSESDecisionTable 1999.04.29 15:43:21 Executing EqualFrequencyScaler... Parameters = {MODE=Save; MASK=T; FILENAME=c:\temp\backupcuts0.txt; INTERVALS=3} 1999.04.29 15:43:21 Output = RSESDecisionTable 1999.04.29 15:43:21 Executing JohnsonReducer... Parameters = {DISCERNIBILITY=Object; MODULO.DECISION=T; SELECTION=All; TYPES=F; PRECOMPUTE=F; APPROXIMATE=F} 1999.04.29 15:43:21 Output = RSESReducts #Members = 8 #Children = 1 1999.04.29 15:43:21 Executing RSESRuleGenerator... Parameters = {} 1999.04.29 15:43:21 Output = RSESRules #Members = 13 1999.04.29 15:43:21 Executing MyRuleFilter... Parameters = {FILTERING=1; CONNECTIVE=Or; SUPPORT.RHS.LOWER=0; SUPPORT.RHS.UPPER=1; INVERT=F} 1999.04.29 15:43:21 Output = RSESRules #Members = 8 1999.04.29 15:43:21 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=F; FILENAME=c:\temp\cuts0.txt} 1999.04.29 15:43:21 Output = RSESDecisionTable 1999.04.29 15:43:21 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=T; FILENAME=c:\temp\backupcuts0.txt} 1999.04.29 15:43:21 Output = RSESDecisionTable 1999.04.29 15:43:21 Executing BatchClassifier... Parameters = {CLASSIFIER=StandardVoter; RULES=No name; TOLERANCE=1.0; SPECIFIC=F; VOTING=Support; NORMALIZATION=Firing; FALLBACK=T; FALLBACK.CLASS=Iris-setosa; FALLBACK.CERTAINTY=0.33; MULTIPLE=Best; LOG=T; LOG.FILENAME=c:\temp\log0.txt; LOG.VERBOSE=T; CALIBRATION=F; ROC=F} 1999.04.29 15:43:21 Output = BatchClassification | 0 1 2 | ----------------------------------------------------- 0 | 3 0 0 | 100.0% 1 | 0 7 0 | 100.0% 2 | 0 0 5 | 100.0% ----------------------------------------------------- | 100.0% 100.0% 100.0% | 100.0% 1999.04.29 15:43:21 Executing BROrthogonalScaler... Parameters = {MODE=Save; MASK=F; FILENAME=c:\temp\cuts1.txt; APPROXIMATE=F} 1999.04.29 15:43:21 Output = RSESDecisionTable 1999.04.29 15:43:21 Executing EqualFrequencyScaler... Parameters = {MODE=Save; MASK=T; FILENAME=c:\temp\backupcuts1.txt; INTERVALS=3} 1999.04.29 15:43:21 Output = RSESDecisionTable 1999.04.29 15:43:21 Executing JohnsonReducer... Parameters = {DISCERNIBILITY=Object; MODULO.DECISION=T; SELECTION=All; TYPES=F; PRECOMPUTE=F; APPROXIMATE=F} 1999.04.29 15:43:21 Output = RSESReducts #Members = 8 #Children = 1 1999.04.29 15:43:21 Executing RSESRuleGenerator... Parameters = {} 1999.04.29 15:43:21 Output = RSESRules #Members = 13 1999.04.29 15:43:21 Executing MyRuleFilter... Parameters = {FILTERING=1; CONNECTIVE=Or; SUPPORT.RHS.LOWER=0; SUPPORT.RHS.UPPER=1; INVERT=F} 1999.04.29 15:43:21 Output = RSESRules #Members = 8 1999.04.29 15:43:21 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=F; FILENAME=c:\temp\cuts1.txt} 1999.04.29 15:43:21 Output = RSESDecisionTable 1999.04.29 15:43:21 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=T; FILENAME=c:\temp\backupcuts1.txt} 1999.04.29 15:43:21 Output = RSESDecisionTable 1999.04.29 15:43:21 Executing BatchClassifier... Parameters = {CLASSIFIER=StandardVoter; RULES=No name; TOLERANCE=1.0; SPECIFIC=F; VOTING=Support; NORMALIZATION=Firing; FALLBACK=T; FALLBACK.CLASS=Iris-setosa; FALLBACK.CERTAINTY=0.33; MULTIPLE=Best; LOG=T; LOG.FILENAME=c:\temp\log1.txt; LOG.VERBOSE=T; CALIBRATION=F; ROC=F} 1999.04.29 15:43:21 Output = BatchClassification | 0 1 2 | ----------------------------------------------------- 0 | 6 0 0 | 100.0% 1 | 0 3 0 | 100.0% 2 | 0 0 6 | 100.0% ----------------------------------------------------- | 100.0% 100.0% 100.0% | 100.0% 1999.04.29 15:43:21 Executing BROrthogonalScaler... Parameters = {MODE=Save; MASK=F; FILENAME=c:\temp\cuts2.txt; APPROXIMATE=F} 1999.04.29 15:43:21 Output = RSESDecisionTable 1999.04.29 15:43:21 Executing EqualFrequencyScaler... Parameters = {MODE=Save; MASK=T; FILENAME=c:\temp\backupcuts2.txt; INTERVALS=3} 1999.04.29 15:43:21 Output = RSESDecisionTable 1999.04.29 15:43:21 Executing JohnsonReducer... Parameters = {DISCERNIBILITY=Object; MODULO.DECISION=T; SELECTION=All; TYPES=F; PRECOMPUTE=F; APPROXIMATE=F} 1999.04.29 15:43:21 Output = RSESReducts #Members = 7 #Children = 1 1999.04.29 15:43:21 Executing RSESRuleGenerator... Parameters = {} 1999.04.29 15:43:21 Output = RSESRules #Members = 12 1999.04.29 15:43:21 Executing MyRuleFilter... Parameters = {FILTERING=1; CONNECTIVE=Or; SUPPORT.RHS.LOWER=0; SUPPORT.RHS.UPPER=1; INVERT=F} 1999.04.29 15:43:21 Output = RSESRules #Members = 8 1999.04.29 15:43:21 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=F; FILENAME=c:\temp\cuts2.txt} 1999.04.29 15:43:21 Output = RSESDecisionTable 1999.04.29 15:43:21 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=T; FILENAME=c:\temp\backupcuts2.txt} 1999.04.29 15:43:21 Output = RSESDecisionTable 1999.04.29 15:43:21 Executing BatchClassifier... Parameters = {CLASSIFIER=StandardVoter; RULES=No name; TOLERANCE=1.0; SPECIFIC=F; VOTING=Support; NORMALIZATION=Firing; FALLBACK=T; FALLBACK.CLASS=Iris-setosa; FALLBACK.CERTAINTY=0.33; MULTIPLE=Best; LOG=T; LOG.FILENAME=c:\temp\log2.txt; LOG.VERBOSE=T; CALIBRATION=F; ROC=F} 1999.04.29 15:43:21 Output = BatchClassification | 0 1 2 | ----------------------------------------------------- 0 | 4 0 0 | 100.0% 1 | 0 4 0 | 100.0% 2 | 0 0 7 | 100.0% ----------------------------------------------------- | 100.0% 100.0% 100.0% | 100.0% 1999.04.29 15:43:21 Executing BROrthogonalScaler... Parameters = {MODE=Save; MASK=F; FILENAME=c:\temp\cuts3.txt; APPROXIMATE=F} 1999.04.29 15:43:21 Output = RSESDecisionTable 1999.04.29 15:43:21 Executing EqualFrequencyScaler... Parameters = {MODE=Save; MASK=T; FILENAME=c:\temp\backupcuts3.txt; INTERVALS=3} 1999.04.29 15:43:21 Output = RSESDecisionTable 1999.04.29 15:43:21 Executing JohnsonReducer... Parameters = {DISCERNIBILITY=Object; MODULO.DECISION=T; SELECTION=All; TYPES=F; PRECOMPUTE=F; APPROXIMATE=F} 1999.04.29 15:43:22 Output = RSESReducts #Members = 7 #Children = 1 1999.04.29 15:43:22 Executing RSESRuleGenerator... Parameters = {} 1999.04.29 15:43:22 Output = RSESRules #Members = 10 1999.04.29 15:43:22 Executing MyRuleFilter... Parameters = {FILTERING=1; CONNECTIVE=Or; SUPPORT.RHS.LOWER=0; SUPPORT.RHS.UPPER=1; INVERT=F} 1999.04.29 15:43:22 Output = RSESRules #Members = 9 1999.04.29 15:43:22 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=F; FILENAME=c:\temp\cuts3.txt} 1999.04.29 15:43:22 Output = RSESDecisionTable 1999.04.29 15:43:22 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=T; FILENAME=c:\temp\backupcuts3.txt} 1999.04.29 15:43:22 Output = RSESDecisionTable 1999.04.29 15:43:22 Executing BatchClassifier... Parameters = {CLASSIFIER=StandardVoter; RULES=No name; TOLERANCE=1.0; SPECIFIC=F; VOTING=Support; NORMALIZATION=Firing; FALLBACK=T; FALLBACK.CLASS=Iris-setosa; FALLBACK.CERTAINTY=0.33; MULTIPLE=Best; LOG=T; LOG.FILENAME=c:\temp\log3.txt; LOG.VERBOSE=T; CALIBRATION=F; ROC=F} 1999.04.29 15:43:22 Output = BatchClassification | 0 1 2 | ----------------------------------------------------- 0 | 5 0 0 | 100.0% 1 | 0 6 1 | 85.71428% 2 | 0 0 3 | 100.0% ----------------------------------------------------- | 100.0% 100.0% 75.0% | 93.33333% 1999.04.29 15:43:22 Executing BROrthogonalScaler... Parameters = {MODE=Save; MASK=F; FILENAME=c:\temp\cuts4.txt; APPROXIMATE=F} 1999.04.29 15:43:22 Output = RSESDecisionTable 1999.04.29 15:43:22 Executing EqualFrequencyScaler... Parameters = {MODE=Save; MASK=T; FILENAME=c:\temp\backupcuts4.txt; INTERVALS=3} 1999.04.29 15:43:22 Output = RSESDecisionTable 1999.04.29 15:43:22 Executing JohnsonReducer... Parameters = {DISCERNIBILITY=Object; MODULO.DECISION=T; SELECTION=All; TYPES=F; PRECOMPUTE=F; APPROXIMATE=F} 1999.04.29 15:43:22 Output = RSESReducts #Members = 7 #Children = 1 1999.04.29 15:43:22 Executing RSESRuleGenerator... Parameters = {} 1999.04.29 15:43:22 Output = RSESRules #Members = 10 1999.04.29 15:43:22 Executing MyRuleFilter... Parameters = {FILTERING=1; CONNECTIVE=Or; SUPPORT.RHS.LOWER=0; SUPPORT.RHS.UPPER=1; INVERT=F} 1999.04.29 15:43:22 Output = RSESRules #Members = 9 1999.04.29 15:43:22 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=F; FILENAME=c:\temp\cuts4.txt} 1999.04.29 15:43:22 Output = RSESDecisionTable 1999.04.29 15:43:22 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=T; FILENAME=c:\temp\backupcuts4.txt} 1999.04.29 15:43:22 Output = RSESDecisionTable 1999.04.29 15:43:22 Executing BatchClassifier... Parameters = {CLASSIFIER=StandardVoter; RULES=No name; TOLERANCE=1.0; SPECIFIC=F; VOTING=Support; NORMALIZATION=Firing; FALLBACK=T; FALLBACK.CLASS=Iris-setosa; FALLBACK.CERTAINTY=0.33; MULTIPLE=Best; LOG=T; LOG.FILENAME=c:\temp\log4.txt; LOG.VERBOSE=T; CALIBRATION=F; ROC=F} 1999.04.29 15:43:22 Output = BatchClassification | 0 1 2 | ----------------------------------------------------- 0 | 4 0 0 | 100.0% 1 | 0 4 1 | 80.0% 2 | 0 0 6 | 100.0% ----------------------------------------------------- | 100.0% 100.0% 85.71428% | 93.33333% 1999.04.29 15:43:22 Executing BROrthogonalScaler... Parameters = {MODE=Save; MASK=F; FILENAME=c:\temp\cuts5.txt; APPROXIMATE=F} 1999.04.29 15:43:22 Output = RSESDecisionTable 1999.04.29 15:43:22 Executing EqualFrequencyScaler... Parameters = {MODE=Save; MASK=T; FILENAME=c:\temp\backupcuts5.txt; INTERVALS=3} 1999.04.29 15:43:22 Output = RSESDecisionTable 1999.04.29 15:43:22 Executing JohnsonReducer... Parameters = {DISCERNIBILITY=Object; MODULO.DECISION=T; SELECTION=All; TYPES=F; PRECOMPUTE=F; APPROXIMATE=F} 1999.04.29 15:43:22 Output = RSESReducts #Members = 7 #Children = 1 1999.04.29 15:43:22 Executing RSESRuleGenerator... Parameters = {} 1999.04.29 15:43:22 Output = RSESRules #Members = 9 1999.04.29 15:43:22 Executing MyRuleFilter... Parameters = {FILTERING=1; CONNECTIVE=Or; SUPPORT.RHS.LOWER=0; SUPPORT.RHS.UPPER=1; INVERT=F} 1999.04.29 15:43:22 Output = RSESRules #Members = 7 1999.04.29 15:43:22 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=F; FILENAME=c:\temp\cuts5.txt} 1999.04.29 15:43:22 Output = RSESDecisionTable 1999.04.29 15:43:22 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=T; FILENAME=c:\temp\backupcuts5.txt} 1999.04.29 15:43:22 Output = RSESDecisionTable 1999.04.29 15:43:22 Executing BatchClassifier... Parameters = {CLASSIFIER=StandardVoter; RULES=No name; TOLERANCE=1.0; SPECIFIC=F; VOTING=Support; NORMALIZATION=Firing; FALLBACK=T; FALLBACK.CLASS=Iris-setosa; FALLBACK.CERTAINTY=0.33; MULTIPLE=Best; LOG=T; LOG.FILENAME=c:\temp\log5.txt; LOG.VERBOSE=T; CALIBRATION=F; ROC=F} 1999.04.29 15:43:22 Output = BatchClassification | 0 1 2 | ----------------------------------------------------- 0 | 6 0 0 | 100.0% 1 | 0 3 1 | 75.0% 2 | 0 1 4 | 80.0% ----------------------------------------------------- | 100.0% 75.0% 80.0% | 86.66666% 1999.04.29 15:43:22 Executing BROrthogonalScaler... Parameters = {MODE=Save; MASK=F; FILENAME=c:\temp\cuts6.txt; APPROXIMATE=F} 1999.04.29 15:43:22 Output = RSESDecisionTable 1999.04.29 15:43:22 Executing EqualFrequencyScaler... Parameters = {MODE=Save; MASK=T; FILENAME=c:\temp\backupcuts6.txt; INTERVALS=3} 1999.04.29 15:43:22 Output = RSESDecisionTable 1999.04.29 15:43:22 Executing JohnsonReducer... Parameters = {DISCERNIBILITY=Object; MODULO.DECISION=T; SELECTION=All; TYPES=F; PRECOMPUTE=F; APPROXIMATE=F} 1999.04.29 15:43:22 Output = RSESReducts #Members = 5 #Children = 1 1999.04.29 15:43:22 Executing RSESRuleGenerator... Parameters = {} 1999.04.29 15:43:22 Output = RSESRules #Members = 11 1999.04.29 15:43:22 Executing MyRuleFilter... Parameters = {FILTERING=1; CONNECTIVE=Or; SUPPORT.RHS.LOWER=0; SUPPORT.RHS.UPPER=1; INVERT=F} 1999.04.29 15:43:22 Output = RSESRules #Members = 9 1999.04.29 15:43:22 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=F; FILENAME=c:\temp\cuts6.txt} 1999.04.29 15:43:22 Output = RSESDecisionTable 1999.04.29 15:43:22 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=T; FILENAME=c:\temp\backupcuts6.txt} 1999.04.29 15:43:22 Output = RSESDecisionTable 1999.04.29 15:43:22 Executing BatchClassifier... Parameters = {CLASSIFIER=StandardVoter; RULES=No name; TOLERANCE=1.0; SPECIFIC=F; VOTING=Support; NORMALIZATION=Firing; FALLBACK=T; FALLBACK.CLASS=Iris-setosa; FALLBACK.CERTAINTY=0.33; MULTIPLE=Best; LOG=T; LOG.FILENAME=c:\temp\log6.txt; LOG.VERBOSE=T; CALIBRATION=F; ROC=F} 1999.04.29 15:43:22 Output = BatchClassification | 0 1 2 | ----------------------------------------------------- 0 | 6 0 0 | 100.0% 1 | 0 6 0 | 100.0% 2 | 0 1 2 | 66.66667% ----------------------------------------------------- | 100.0% 85.71428% 100.0% | 93.33333% 1999.04.29 15:43:22 Executing BROrthogonalScaler... Parameters = {MODE=Save; MASK=F; FILENAME=c:\temp\cuts7.txt; APPROXIMATE=F} 1999.04.29 15:43:23 Output = RSESDecisionTable 1999.04.29 15:43:23 Executing EqualFrequencyScaler... Parameters = {MODE=Save; MASK=T; FILENAME=c:\temp\backupcuts7.txt; INTERVALS=3} 1999.04.29 15:43:23 Output = RSESDecisionTable 1999.04.29 15:43:23 Executing JohnsonReducer... Parameters = {DISCERNIBILITY=Object; MODULO.DECISION=T; SELECTION=All; TYPES=F; PRECOMPUTE=F; APPROXIMATE=F} 1999.04.29 15:43:23 Output = RSESReducts #Members = 8 #Children = 1 1999.04.29 15:43:23 Executing RSESRuleGenerator... Parameters = {} 1999.04.29 15:43:23 Output = RSESRules #Members = 13 1999.04.29 15:43:23 Executing MyRuleFilter... Parameters = {FILTERING=1; CONNECTIVE=Or; SUPPORT.RHS.LOWER=0; SUPPORT.RHS.UPPER=1; INVERT=F} 1999.04.29 15:43:23 Output = RSESRules #Members = 8 1999.04.29 15:43:23 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=F; FILENAME=c:\temp\cuts7.txt} 1999.04.29 15:43:23 Output = RSESDecisionTable 1999.04.29 15:43:23 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=T; FILENAME=c:\temp\backupcuts7.txt} 1999.04.29 15:43:23 Output = RSESDecisionTable 1999.04.29 15:43:23 Executing BatchClassifier... Parameters = {CLASSIFIER=StandardVoter; RULES=No name; TOLERANCE=1.0; SPECIFIC=F; VOTING=Support; NORMALIZATION=Firing; FALLBACK=T; FALLBACK.CLASS=Iris-setosa; FALLBACK.CERTAINTY=0.33; MULTIPLE=Best; LOG=T; LOG.FILENAME=c:\temp\log7.txt; LOG.VERBOSE=T; CALIBRATION=F; ROC=F} 1999.04.29 15:43:23 Output = BatchClassification | 0 1 2 | ----------------------------------------------------- 0 | 4 0 0 | 100.0% 1 | 0 4 0 | 100.0% 2 | 0 0 7 | 100.0% ----------------------------------------------------- | 100.0% 100.0% 100.0% | 100.0% 1999.04.29 15:43:23 Executing BROrthogonalScaler... Parameters = {MODE=Save; MASK=F; FILENAME=c:\temp\cuts8.txt; APPROXIMATE=F} 1999.04.29 15:43:23 Output = RSESDecisionTable 1999.04.29 15:43:23 Executing EqualFrequencyScaler... Parameters = {MODE=Save; MASK=T; FILENAME=c:\temp\backupcuts8.txt; INTERVALS=3} 1999.04.29 15:43:23 Output = RSESDecisionTable 1999.04.29 15:43:23 Executing JohnsonReducer... Parameters = {DISCERNIBILITY=Object; MODULO.DECISION=T; SELECTION=All; TYPES=F; PRECOMPUTE=F; APPROXIMATE=F} 1999.04.29 15:43:23 Output = RSESReducts #Members = 6 #Children = 1 1999.04.29 15:43:23 Executing RSESRuleGenerator... Parameters = {} 1999.04.29 15:43:23 Output = RSESRules #Members = 12 1999.04.29 15:43:23 Executing MyRuleFilter... Parameters = {FILTERING=1; CONNECTIVE=Or; SUPPORT.RHS.LOWER=0; SUPPORT.RHS.UPPER=1; INVERT=F} 1999.04.29 15:43:23 Output = RSESRules #Members = 11 1999.04.29 15:43:23 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=F; FILENAME=c:\temp\cuts8.txt} 1999.04.29 15:43:23 Output = RSESDecisionTable 1999.04.29 15:43:23 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=T; FILENAME=c:\temp\backupcuts8.txt} 1999.04.29 15:43:23 Output = RSESDecisionTable 1999.04.29 15:43:23 Executing BatchClassifier... Parameters = {CLASSIFIER=StandardVoter; RULES=No name; TOLERANCE=1.0; SPECIFIC=F; VOTING=Support; NORMALIZATION=Firing; FALLBACK=T; FALLBACK.CLASS=Iris-setosa; FALLBACK.CERTAINTY=0.33; MULTIPLE=Best; LOG=T; LOG.FILENAME=c:\temp\log8.txt; LOG.VERBOSE=T; CALIBRATION=F; ROC=F} 1999.04.29 15:43:23 Output = BatchClassification | 0 1 2 | ----------------------------------------------------- 0 | 8 0 0 | 100.0% 1 | 0 4 0 | 100.0% 2 | 0 0 3 | 100.0% ----------------------------------------------------- | 100.0% 100.0% 100.0% | 100.0% 1999.04.29 15:43:23 Executing BROrthogonalScaler... Parameters = {MODE=Save; MASK=F; FILENAME=c:\temp\cuts9.txt; APPROXIMATE=F} 1999.04.29 15:43:23 Output = RSESDecisionTable 1999.04.29 15:43:23 Executing EqualFrequencyScaler... Parameters = {MODE=Save; MASK=T; FILENAME=c:\temp\backupcuts9.txt; INTERVALS=3} 1999.04.29 15:43:23 Output = RSESDecisionTable 1999.04.29 15:43:23 Executing JohnsonReducer... Parameters = {DISCERNIBILITY=Object; MODULO.DECISION=T; SELECTION=All; TYPES=F; PRECOMPUTE=F; APPROXIMATE=F} 1999.04.29 15:43:23 Output = RSESReducts #Members = 7 #Children = 1 1999.04.29 15:43:23 Executing RSESRuleGenerator... Parameters = {} 1999.04.29 15:43:23 Output = RSESRules #Members = 12 1999.04.29 15:43:23 Executing MyRuleFilter... Parameters = {FILTERING=1; CONNECTIVE=Or; SUPPORT.RHS.LOWER=0; SUPPORT.RHS.UPPER=1; INVERT=F} 1999.04.29 15:43:23 Output = RSESRules #Members = 9 1999.04.29 15:43:23 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=F; FILENAME=c:\temp\cuts9.txt} 1999.04.29 15:43:23 Output = RSESDecisionTable 1999.04.29 15:43:23 Executing OrthogonalFileScaler... Parameters = {MODE=Load; MASK=T; FILENAME=c:\temp\backupcuts9.txt} 1999.04.29 15:43:23 Output = RSESDecisionTable 1999.04.29 15:43:23 Executing BatchClassifier... Parameters = {CLASSIFIER=StandardVoter; RULES=No name; TOLERANCE=1.0; SPECIFIC=F; VOTING=Support; NORMALIZATION=Firing; FALLBACK=T; FALLBACK.CLASS=Iris-setosa; FALLBACK.CERTAINTY=0.33; MULTIPLE=Best; LOG=T; LOG.FILENAME=c:\temp\log9.txt; LOG.VERBOSE=T; CALIBRATION=F; ROC=F} 1999.04.29 15:43:23 Output = BatchClassification | 0 1 2 | ----------------------------------------------------- 0 | 4 0 0 | 100.0% 1 | 0 6 0 | 100.0% 2 | 0 1 4 | 80.0% ----------------------------------------------------- | 100.0% 85.71428% 100.0% | 93.33333% Accuracy.Mean = 0.96 Accuracy.Median = 0.966667 Accuracy.StdDev = 0.046614 Accuracy.Minimum = 0.866667 Accuracy.Maximum = 1.0