jcolibri.method.retrieve.DiverseByMedianRetrieval
Class ExpertClerkMedianScoring

java.lang.Object
  extended by jcolibri.method.retrieve.DiverseByMedianRetrieval.ExpertClerkMedianScoring

public class ExpertClerkMedianScoring
extends java.lang.Object

ExpertClerk Median algorithm. This algorithm chooses the first case that is closed to the median of cases. Then the remaining are selected taking into account negative and possitive characteristics. A characteristic is an attribute that exceeds a predefined threshold. It is positive if is greater thatn the value of the median. And negative otherwise. The number of positive plus the negative characteristics is used to rank the cases and obtain the following k-1 cases.

See:

H. Shimazu. ExpertClerk: A Conversational Case-Based Reasoning Tool for Developing Salesclerk Agents in E-Commerce Webshops. Artif. Intell. Rev., 18(3-4):223-244, 2002.

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

Constructor Summary
ExpertClerkMedianScoring()
           
 
Method Summary
static java.util.Collection<RetrievalResult> getDiverseByMedian(java.util.Collection<CBRCase> cases, NNConfig simConfig, java.util.HashMap<Attribute,java.lang.Double> thresholds)
          Returns diverse cases using the ExpertClerk median method.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

ExpertClerkMedianScoring

public ExpertClerkMedianScoring()
Method Detail

getDiverseByMedian

public static java.util.Collection<RetrievalResult> getDiverseByMedian(java.util.Collection<CBRCase> cases,
                                                                       NNConfig simConfig,
                                                                       java.util.HashMap<Attribute,java.lang.Double> thresholds)
Returns diverse cases using the ExpertClerk median method.

Parameters:
cases - to retrieve from
simConfig - is the nn configuration
thresholds - to obtain the characteristics
Returns:
a collection of cases

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