jcolibri.method.revise.classification
Interface ClassificationOracle

All Known Implementing Classes:
BasicClassificationOracle

public interface ClassificationOracle

Interface that defines functions to decide if a query has been correctly classified or not and to assign a cost to incorrectly classified queries.

Author:
Derek Bridge, Lisa Cummins 16/05/07

Method Summary
 double getPredictionCost(CBRQuery query, CBRCaseBase caseBase, KNNClassificationConfig knnConfig)
          Calculates the cost of the prediction made by the given case-base and similarity configuration for the solution of the query.
 double getPredictionCost(ClassificationSolution predictedSolution, CBRCase testCase)
          Calculates the cost of the given solution as a prediction for the solution of the given case.
 double getPredictionCost(ClassificationSolution predictedSolution, ClassificationSolution correctSolution)
          Calculates the cost of the given test solution while bein compared to the given correct solution.
 boolean isCorrectPrediction(CBRQuery query, CBRCaseBase caseBase, KNNClassificationConfig knnConfig)
          Checks if the query is correctly classified by the given case-base and similarity configuration.
 boolean isCorrectPrediction(ClassificationSolution predictedSolution, CBRCase testCase)
          Checks if the predicted solution is the correct solution for the given test case.
 boolean isCorrectPrediction(ClassificationSolution predictedSolution, ClassificationSolution correctSolution)
          Checks if the predicted solution and the correct solution are the same.
 

Method Detail

isCorrectPrediction

boolean isCorrectPrediction(ClassificationSolution predictedSolution,
                            CBRCase testCase)
Checks if the predicted solution is the correct solution for the given test case.

Parameters:
predictedSolution - the predicted solution.
testCase - the test case (query and correct solution).
Returns:
true if the predicted solution is the correct solution for the given test case, false if not.

isCorrectPrediction

boolean isCorrectPrediction(ClassificationSolution predictedSolution,
                            ClassificationSolution correctSolution)
Checks if the predicted solution and the correct solution are the same.

Parameters:
predictedSolution - the predicted solution.
correctSolution - the correct solution.
Returns:
true if the predicted solution and the correct solution are the same, false if not.

isCorrectPrediction

boolean isCorrectPrediction(CBRQuery query,
                            CBRCaseBase caseBase,
                            KNNClassificationConfig knnConfig)
Checks if the query is correctly classified by the given case-base and similarity configuration.

Parameters:
query - the query to be tested.
caseBase - the case base to use to find the predicted solution.
knnConfig - the similarity configuration.
Returns:
true if the query is correctly classified by the given case-base and similarity configuration, fasle otherwise.

getPredictionCost

double getPredictionCost(ClassificationSolution predictedSolution,
                         CBRCase testCase)
Calculates the cost of the given solution as a prediction for the solution of the given case.

Parameters:
predictedSolution - the predicted solution.
testCase - the test case (query and correct solution).
Returns:
the cost of the prediction made.

getPredictionCost

double getPredictionCost(ClassificationSolution predictedSolution,
                         ClassificationSolution correctSolution)
Calculates the cost of the given test solution while bein compared to the given correct solution.

Parameters:
predictedSolution - the predicted solution.
correctSolution - the correct solution.
Returns:
the cost of the prediction made.

getPredictionCost

double getPredictionCost(CBRQuery query,
                         CBRCaseBase caseBase,
                         KNNClassificationConfig knnConfig)
Calculates the cost of the prediction made by the given case-base and similarity configuration for the solution of the query.

Parameters:
query - the query to be tested.
caseBase - the case base to use to find the predicted solution.
knnConfig - the similarity configuration.
Returns:
the cost of the prediction made.

GAIA - Group for Artificial Intelligence Applications
http://gaia.fdi.ucm.es