Publications
|
Les dépendances fonctionnelles pour la sélection de vues dans les cubes de données
Authors : Eve Garnaud, Sofian Maabout, Mohamed Mosbah
Journal : Journal of Decision Systems, Volume 21
Web : http://www.tandfonline.com/doi/abs/10.1080/12460125.2012.677645
OLAP query processing assumes two seemingly contradictory requirements: on one hand query processing should be fast (thus, queries pre-precomputation) and on another hand queries are assumed to be submitted in an ad hoc manner making workload usage for optimization sometimes not effective (we cannot materialize an infinite number of queries). Thus, in this paper we address the specific case of data cubes and a special class of queries corresponding somehow to report queries. We present techniques to select the queries to be materialized with the help of exact and approximate functional dependencies (FD’s). In contrast to most of the previous approaches, we take a user instead of a DBA point of view: rather than fixing the available storage space and minimizing queries cost, we fix the maximal queries cost and we minimize the required storage space. The conducted experiments show the effectiveness of our results. |
|
Publications
|
Multi-strategies Ontology Alignment Aggregated by AHP
Authors : Fuqi Song; Gregory Zacharewicz; David Chen
Conf : 16th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems(KES), San Sebastian : Spain (2012)
Web : http://hal.archives-ouvertes.fr/hal-00727155
Ontology alignment is a critical issue for semantic integration and data interoperability. Many basic matching techniques have been proposed, the crucial point is how to combine different techniques to adapt the diverse sources of ontology. In this paper, we proposed to use three matchers with multiple strategies from perspective of label, semantic and structural aspects of ontology. A rule-based pre-process for identifying the core word with part-of-speech (POS) tagging is carried out before matching. Method Analytic Hierarchy Process (AHP) is applied to aggregate different matching results dynamically and automatically. The intensity of importance is determined by three similarity indicators, which reflect the essential features of source ontologies. This proposed approach shows strong matching ability, which obtained high precision and outstanding evaluation results in benchmarking experiment. |
Publications
|
An Ontology-Driven Framework towards Building Enterprise Semantic Information Layer
Authors : Fuqi Song; Gregory Zacharewicz; David Chen
Journal : Advanced Engineering Informatics, 2013, 27 (1), pp. 38-50
Web : http://hal.archives-ouvertes.fr/hal-00727155 |
|
|
|