ibspan.tss.um.filtering
Class FilteringAlgorithm

java.lang.Object
  ibspan.tss.um.filtering.FilteringAlgorithm

public class FilteringAlgorithm
extends Object

The class realize personalization of restaraunts list using filtering algorithm. The algorithm exploite the user myProfile in the following order:

1. For whole set of results rate each restaurant identified by root node R accordingly to the definition of temperature of potential interest (see [1] for details).
   1. (a) Calculate statistics (number of occurence) for reflection of each resource in R-subgraph.
   2. Compute temperature for each R-subgraph.
2. Compute level L for rated restaurants.
3. Throw away restaurants with temperature below the level L.
4. Sort restaurants descendingly according to the assigned temperatures (results will be sorted by view transforming service)

The algorithm is described in detail in [1], it corrects solutions presented in [2] and is inspired by works [3,4].

1. "Exploitation of User Profile in Travel Support System -- Ontology-based Filtering Algorithm", Technical Report, Maciej Gawinecki, August 10, 2006 (attached)
2. "Modelling User on the Basis of Interactions with a WWW Based System", Maciej Gawinecki, Adam Mickiewicz University, Poznan. 2005.
3. "Ontology-based Stereotyping in a Travel Support System." Maciej Gawinecki, Mateusz Kruszyk, and Marcin Paprzycki. In Proceedings of the XXI Fall Meeting of Polish Information Processing Society. 2005.
4. "Utilizing Semantic Web and Software Agents in a Travel Support System". In "Semantic Web Technologies and eBusiness: Virtual Organization and Business Process Automation." Maria Ganzha, Maciej Gawinecki, Marcin Paprzycki, Rafal Gasiorowski, Szymon Pisarek, and Wawrzyniec Hyska. Editted by A. F. Salam and Jason Stevens. Idea Publishing Group. 2006.

Author:

Maciej Gawinecki

Field Summary

static double	ALPHA Predefined constant configures influence of normalized probability on temperature of single opinion.
static double	BETA Predefined constant configures influence of UserContext onto temperature of single opinion.

Constructor Summary

FilteringAlgorithm(UserProfile profile, OntModel mDomain)
Constructs FilteringAlgorithm.

Method Summary

Memory

filter(Memory results, Map context) Filters given results according to the user myProfile and the given context of user behaviour.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

ALPHA

public static final double ALPHA

Predefined constant configures influence of normalized probability on temperature of single opinion.

See Also:

Constant Field Values

BETA

public static final double BETA

Predefined constant configures influence of UserContext onto temperature of single opinion.

See Also:

Constant Field Values

Constructor Detail

FilteringAlgorithm

public FilteringAlgorithm(UserProfile profile,
                          OntModel mDomain)

Constructs FilteringAlgorithm.

Parameters:

myProfile - user profile which will be used a base for filtering

mDomain - ontology of domain, to which references results which are to be filtered, i.e. ont

Throws:

UserModellingDomainAccessException

com.ibm.adtech.jastor.JastorException

Method Detail

filter

public Memory filter(Memory results,
                     Map context)
              throws AnonymousResourceException

Filters given results according to the user myProfile and the given context of user behaviour. In fact, it removes from results Memory those who are rated too low.

Parameters:

results - results to be filtered

context - context of user behaviour

Returns:

the memory without not interesting results

Throws:

AnonymousResourceException