org.processmining.analysis.decisionmining
Class DecisionAnalyser

java.lang.Object
  extended by org.processmining.analysis.decisionmining.DecisionAnalyser
Direct Known Subclasses:
J48Analyser

public abstract class DecisionAnalyser
extends java.lang.Object

Class holding the machine learning settings and performing the actual data mining tasks (with the help of the weka library).

Author:
arozinat (a.rozinat@tm.tue.nl)

Field Summary
protected  weka.classifiers.Classifier myClassifier
          The data mining classifier to be used for analysis.
 
Constructor Summary
DecisionAnalyser()
           
 
Method Summary
 void analyse(java.util.List<DecisionPoint> decisionPoints, DecisionMiningLogReader log, HLPetriNet highLevelPN)
          Analyses the given list of decision points according to the context specified.
protected  javax.swing.JPanel createEvaluationVisualization(weka.core.Instances data)
          Creates an evaluation overview of the built classifier.
static javax.swing.JPanel createMessagePanel(java.lang.String message)
          Helper method creating an empty panel containing the given feedback message for the user.
protected abstract  javax.swing.JPanel createResultVisualization()
          Creates some result visualization of the result for the current classification problem.
abstract  java.lang.String getDescription()
          Provide a description of this algorithm to be displayed in the algorithm view.
abstract  javax.swing.JPanel getParametersPanel()
          Creates a GUI panel containing the parameters that are available for this type of decision analyser (i.e., the used algorithm).
protected abstract  void initClassifier()
          Initializes data mining classifier to be used for analysis.
protected abstract  void redrawResultVisualization(javax.swing.JPanel panel)
          Invokes the redraw of the given result visualization.
abstract  java.lang.String toString()
          Specify the name of this algorithm to be displayed in the combo box of the algorithm view.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait
 

Field Detail

myClassifier

protected weka.classifiers.Classifier myClassifier
The data mining classifier to be used for analysis.

Constructor Detail

DecisionAnalyser

public DecisionAnalyser()
Method Detail

analyse

public void analyse(java.util.List<DecisionPoint> decisionPoints,
                    DecisionMiningLogReader log,
                    HLPetriNet highLevelPN)
Analyses the given list of decision points according to the context specified. Furthermore, the context is provided with some visualization of the analysis result.

Parameters:
decisionPoints - the list of decision points to be analysed
log - the log to be analysed
highLevelPN - the simulation model to export discovered data dependencies

createMessagePanel

public static javax.swing.JPanel createMessagePanel(java.lang.String message)
Helper method creating an empty panel containing the given feedback message for the user.

Parameters:
message - the message to be displayed for the user
Returns:
the panel to be displayed as analysis result for the current decision point

createEvaluationVisualization

protected javax.swing.JPanel createEvaluationVisualization(weka.core.Instances data)
Creates an evaluation overview of the built classifier.

Returns:
the panel to be displayed as result evaluation view for the current decision point

toString

public abstract java.lang.String toString()
Specify the name of this algorithm to be displayed in the combo box of the algorithm view.

Overrides:
toString in class java.lang.Object
Returns:
the name of this algorithm

getDescription

public abstract java.lang.String getDescription()
Provide a description of this algorithm to be displayed in the algorithm view.

Returns:
the description of this algorithm

getParametersPanel

public abstract javax.swing.JPanel getParametersPanel()
Creates a GUI panel containing the parameters that are available for this type of decision analyser (i.e., the used algorithm).

Returns:
the parameters panel to be displayed in the algorithm settings of the decsion miner

initClassifier

protected abstract void initClassifier()
Initializes data mining classifier to be used for analysis.


createResultVisualization

protected abstract javax.swing.JPanel createResultVisualization()
Creates some result visualization of the result for the current classification problem.

Returns:
the panel to be displayed as analysis result for the current decision point

redrawResultVisualization

protected abstract void redrawResultVisualization(javax.swing.JPanel panel)
Invokes the redraw of the given result visualization. Note that the visualization is not re-created but only adjusted, e.g., with respect to positioning graphical components which cannot be done before the result panel is actually packed to the frame.

Parameters:
panel - the result visualization to be adjusted