• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 1
  • Tagged with
  • 2
  • 2
  • 2
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

Découverte et exploitation de proportions analogiques dans les bases de données relationnelles / Discovering and exploiting analogical proportions in a relational database context

Correa Beltran, William 18 July 2016 (has links)
Dans cette thèse, nous nous intéressons aux proportions analogiques dans le contexte des bases de données relationnelles. Les proportions analogiques permettent de lier quatre éléments dans une relation du type ''A est à B ce que C est à D''. Par exemple, « Paris est à la France ce que Rome est à l'Italie ». Nous avons étudié le problème de la prédiction de valeurs manquantes dans une base de données en utilisant les proportions analogiques. Un algorithme de classification fondé sur les proportions analogiques a été modifié afin de résoudre ce problème. Puis, nous avons étudié les propriétés des éléments appartenant à l'ensemble d'apprentissage des classificateurs analogiques fréquemment exploités pour calculer la prédiction. Ceci nous a permis de réduire considérablement la taille de cet ensemble par élimination des éléments peu pertinents et par conséquent, de diminuer les temps d'exécution de ces classificateurs. La deuxième partie de la thèse a pour objectif de découvrir de nouveaux patrons basés sur la relation d'analogie, i.e., des parallèles, dans les bases de données. Nous avons montré qu'il est possible d'extraire ces patrons en s'appuyant sur des approches de clustering. Les clusters produits par de telles techniques présentent aussi un intérêt pour l'évaluation de requêtes recherchant des patrons d'analogie dans les bases de données. Dans cette perspective, nous avons proposé d'étendre le langage de requêtes SQL pour pouvoir trouver des quadruplets d'une base de données satisfaisant une proportion analogique. Nous avons proposé différentes stratégies d'évaluation pour de telles requêtes, et avons comparé expérimentalementleurs performances. / In this thesis, we are interested in the notion of analogical proportions in a relational database context. An analogical proportion is a statement of the form “A is to B as C is to D”, expressing that the relation beween A and B is the same as the relation between C and D. For instance, one may say that “Paris is to France as Rome is to Italy”. We studied the problem of imputing missing values in a relational database by means of analogical proportions. A classification algorithm based on analogical proportions has been modified in order to impute missing values. Then, we studied how analogical classifiers work in order to see if their processing could be simplified. We showed how some typeof analogical proportions is more useful than the others when performing classification. We then proposed an algorithm using this information, which allowed us to considerably reduce the size of the training set used by the analogical classificationalgorithm, and hence to reduce its execution time. In the second part of this thesis, we payed a particular attention to the mining of combinations of four tuples bound by an analogical relationship. For doing so, we used several clustering algorithms, and we proposed some modifications to them, in order tomake each obtained cluster represent a set of analogical proportions. Using the results of the clustering algorithms, we studied how to efficiently retrieve the analogical proportions in a database by means of queries. For doing so, we proposed to extend the SQL query language in order to retrieve from a database the quadruples of tuples satisfying an analogical proportion. We proposed severalquery evaluation strategies and experimentally compared their performances.
2

Data Aggregation through Web Service Composition in Smart Camera Networks

Rajapaksage, Jayampathi S 14 December 2010 (has links)
Distributed Smart Camera (DSC) networks are power constrained real-time distributed embedded systems that perform computer vision using multiple cameras. Providing data aggregation techniques that is criti-cal for running complex image processing algorithms on DSCs is a challenging task due to complexity of video and image data. Providing highly desirable SQL APIs for sophisticated query processing in DSC networks is also challenging for similar reasons. Research on DSCs to date have not addressed the above two problems. In this thesis, we develop a novel SOA based middleware framework on a DSC network that uses Distributed OSGi to expose DSC network services as web services. We also develop a novel web service composition scheme that aid in data aggregation and a SQL query interface for DSC net-works that allow sophisticated query processing. We validate our service orchestration concept for data aggregation by providing query primitive for face detection in smart camera network.

Page generated in 0.0421 seconds