jcolibri.extensions.recommendation.navigationByAsking
Class SimilarityInfluence

java.lang.Object
  extended by jcolibri.extensions.recommendation.navigationByAsking.SimilarityInfluence
All Implemented Interfaces:
SelectAttributeMethod

public class SimilarityInfluence
extends java.lang.Object
implements SelectAttributeMethod

Selects the attribute that has the highest infuence on the KNN similarity. The inuence on the similarity can be measured by the expected variance of the similarities of a set of selected cases. This method is not recommended with large case bases.

See:

R. Bergmann. Experience Management: Foundations, Development Methodology, and Internet-Based Applications. Springer-Verlag New York, Inc.,Secaucus, NJ, USA, 2002.

A. Kohlmaier, S. Schmitt, and R. Bergmann. A similarity-based approach to attribute selection in user-adaptive sales dialogs. In D. W. Aha and I. Watson, editors, Proceedings of the 4th International Conference on Case-Based Reasoning, pages 306320, Seattle, Washington, 2001. Springer-Verlag.

S. Schmitt, P. Dopichaj, and P. Domínguez-Marín. Entropy-based vs. similarity-inuenced: Attribute selection methods for dialogs tested on different electronic commerce domains. In S. Craw and A. Preece, editors, Proceedings of the 6th European Conference on Case-Based Reasoning, pages 380-394, Aberdeen, Scotland, 2002. Springer-Verlag.

Version:
1.0
Author:
Juan A. Recio-Garcia, Developed at University College Cork (Ireland) in collaboration with Derek Bridge.

Constructor Summary
SimilarityInfluence(NNConfig simConfig)
          Constructor
 
Method Summary
 Attribute getAttribute(java.util.Collection<CBRCase> cases, CBRQuery query)
          Selects the attribute to be asked
static Attribute getMoreSimVarAttribute(java.util.Collection<CBRCase> cases, CBRQuery query, NNConfig simConfig, boolean init)
          Selects the attribute with more expected influence in the NN scoring.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

SimilarityInfluence

public SimilarityInfluence(NNConfig simConfig)
Constructor

Parameters:
simConfig - is the KNN configuration
Method Detail

getMoreSimVarAttribute

public static Attribute getMoreSimVarAttribute(java.util.Collection<CBRCase> cases,
                                               CBRQuery query,
                                               NNConfig simConfig,
                                               boolean init)
                                        throws ExecutionException
Selects the attribute with more expected influence in the NN scoring.

Parameters:
cases - Set of working cases
query - Query to compare with the cases
simConfig - is the NN similiarity configuration
init - indicates if this is the first time that the algorithm is executed. This way, in following iterations past chosen attributes are not computed.
Returns:
the selected attribute or null if there are not more attributes to ask.
Throws:
ExecutionException

getAttribute

public Attribute getAttribute(java.util.Collection<CBRCase> cases,
                              CBRQuery query)
                       throws ExecutionException
Selects the attribute to be asked

Specified by:
getAttribute in interface SelectAttributeMethod
Parameters:
cases - list of working cases
query - is the current query
Returns:
selected attribute
Throws:
ExecutionException

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