Authors:
Fatma Ezzahra Bousnina
1
;
Mouna Chebbah
2
;
Mohamed Anis Bach Tobji
3
;
Allel Hadjali
4
and
Boutheina Ben Yaghlane
5
Affiliations:
1
Université de Tunis, Institut Supérieur de Gestion and LARODEC, Tunisia
;
2
Université de Tunis, Institut Supérieur de Gestion, LARODEC, Université de Jendouba, Faculté des Sciences Juridiques and Economiques et de Gestion de Jendouba, Tunisia
;
3
Université de Tunis, Institut Supérieur de Gestion, LARODEC, Univ. Manouba and ESEN, Tunisia
;
4
Université de Poitiers, Ecole Nationale Supérieure de Mécanique et de l'Aérotechnique and LIAS, France
;
5
Université de Tunis, Institut Supérieur de Gestion, LARODEC, Université de Carthage and Institut des Hautes Etudes Commerciales, Tunisia
Keyword(s):
Evidence Theory ; Evidential Databases; Top-K Queries; Ranking Intervals; Score function.
Related
Ontology
Subjects/Areas/Topics:
Databases and Information Systems Integration
;
Enterprise Information Systems
;
Query Languages and Query Processing
Abstract:
Uncertain data are obvious in a lot of domains such as sensor networks, multimedia, social media, etc. Top-k queries provide ordered results according to a defined score.
This kind of queries represents an important tool for exploring uncertain data. Most of works cope with certain data and with probabilistic top-k queries.
However, at the best of our knowledge there is no work that exploits the Top-k semantics in the Evidence Theory context.
In this paper, we introduce a new score function suitable for Evidential Data. Since the result of the score function is an interval, we adopt a comparison method for ranking intervals.
Finally we extend the usual semantics/interpretations of top-k queries to the evidential scenario.