Scribd
Upload
33 views

Evaluation of Instances Asset in A Topic Maps Ontology

This document discusses evaluating topics in a topic maps ontology by assigning weights. It proposes assigning partial weights to individual topic attributes, total topic weights based on surrounding topic associations weighted by a coefficient, and normalizing the total weights. Weights can be assigned to different association categories and attributes. Calculating total topic weights allows ranking search results by information value, with more specific weight settings providing better results than defaults.

Uploaded by

tmra-conference
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
33 views

Evaluation of Instances Asset in A Topic Maps Ontology

This document discusses evaluating topics in a topic maps ontology by assigning weights. It proposes assigning partial weights to individual topic attributes, total topic weights based on surrounding topic associations weighted by a coefficient, and normalizing the total weights. Weights can be assigned to different association categories and attributes. Calculating total topic weights allows ranking search results by information value, with more specific weight settings providing better results than defaults.

Uploaded by

tmra-conference
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 12

Evaluation of Instances Asset

in a Topic Maps Ontology

Petra Haluzová
[email protected]

Department of Informatics and Telecommunications


Faculty of Transportation Sciences
Czech Technical University in Prague

TMRA 2010
Information asset of instance (topic)

 Expresses the richness of topic description in the ontology


 The richness of surrounding topics description are taken into account
 Association types between topics are also considered

 Example: Topic Type – dog


 1) Name: Barney
 2) Colour: black

Breed: dachshund
Photo: http://...
?
Associations – owner, vaccination,…
Weights assignment:
Partial weight of topic ci
 Partial weight of topic c = {1, 2,...}; c  N List of attributes and
their weights:
 Expresses the richness of description
Article: 1
of each individual topic Audio recording: 2
 Is equal to the sum of attribute weights Bibliography: 2
Biographical article: 1
 The topic name is an attribute as well Date of birth: 1
 Default setting: 1 for all attribute weights, Date of death: 1
Description: 1
user can change this setting Editorial guidelines: 1
Gallery: 1
Illustration: 1
Libretto: 2
Note: 1
Poster: 1
Premiere date:1
Sound clip: 2
Synopsis: 1
Video recording: 2
Web page: 2
Web site: 3
Weights assignment:
Total topic weight wi
 Total topic weight wi for each individual topic i

wi  ci  k   aij  c j
j

 The partial weight c is calculated for each individual topic in the ontology
 The next part of the total topic weight w is derived from associations with
surrounding topics j
 Coefficient k influences the importance of surrounding topics
 After calculating the total topic weights the results are normalized to
interval (0,1
Weights assignment:
Weights of associations categories aij
 Association weight a   0, 1; a  R
 Three categories: hierarchical, defining, contextual
 Weights setting of all categories at 1 advantages topics which have
great number of associations (independently of association types)

k = 0.2
weights of all attributes: 1
Weights assignment:
Influence of coefficient k
 Coefficient k   0, 1; k  R
 Influences the importance of partial weights of surrounding topics
 The less coefficient value the greater importance is assigned to the
central topic

weights of all attributes: 1


weights of all categories: 1.0
Parameters setting
Results
 Better results are achieved with more specific setting of attribute weights
and category weights than with default setting

 Example:

Weights of attributes in the list above.

k = 0.2

Weights of categories of associations:


defining – 1.0
hierarchical – 0.5
contextual – 0.2
Summary
(repetition is the mother of wisdom)

 The user assigns the weights to all attributes which can occur within a topic
 Default: weight of all attributes 1
 The user divides associations into three categories and assigns the
weights to these categories
 Default: weight of all associations 1
 The total topic weight will be calculated for each individual topic in the
ontology
 These total weights values will be normalized in accordance with the
maximal value
Application
 Information asset points to potential usefulness of information contained in
the topic for the user
 The user’s insight is taken into consideration, if desired
 Topics found during a search in the ontology may be ranked according to
their information asset
 The total topic weights are calculated just once
 The quantification of information asset is used as measure and statistical
data
Scheme of topics interconnection
Matrix notation
Algorithm
Thank you for your attention

You might also like