Design of a learning methode for automatic data extraction / Bouamama, Salim  

Public ISBD Titre : Design of a learning methode for automatic data extraction Type de document : texte imprimé Auteurs : Bouamama, Salim, Acteur ; Abdellah BOUKERRAM, Directeur de thèse ; Christian BLUM Editeur : Setif:UFA Année de publication : 2013 Importance : 1 vol (128 f .) Format : 29 cm Catégories : Informatique Mots-clés : Données de grande dimension Regroupement spectral Apprentissage en utilisant lesgraphes Régularisation sur graphes Marches aléatoires sur graphes Diffusion maps Analyse vidéo du comportement humain Segmentation hiérarchique des flux vidéo Moments de Zernike spatio-temporels Points SIFT Index. décimale : 004 Informatique Résumé : Optimization is a scientific discipline that is concerned with the extraction of optimalsolutions for a problem, among alternatives. Many challenging applications inbusiness, economics,and engineering can be formulated as optimization problems.However, they are often complexand difficult to solve by an exact method within areasonable amount of time.In this thesis we propose a novel approach called population based iterated greedy algorithm in order to efficiently explore and exploit the search space of one of the NP-hard combinatorial optimizationproblems namely the minimumn weigh vertex cover problem. It isa fundamental graph problem with many important real-life applicationssuchas, for example, in wireless communication, circuit design and network flows.An extensive experimental evaluation on a commonly used set of benchmarkinstances shows that our algorithm outperforms current state-of-the-art methods notonly in solution quality but also in computation time. Côte titre : DI/0010 En ligne : http://dspace.univ-setif.dz:8888/jspui/handle/123456789/35 Design of a learning methode for automatic data extraction [texte imprimé] / Bouamama, Salim, Acteur ; Abdellah BOUKERRAM, Directeur de thèse ; Christian BLUM . - [S.l.] : Setif:UFA, 2013 . - 1 vol (128 f .) ; 29 cm. Catégories : Informatique Mots-clés : Données de grande dimension Regroupement spectral Apprentissage en utilisant lesgraphes Régularisation sur graphes Marches aléatoires sur graphes Diffusion maps Analyse vidéo du comportement humain Segmentation hiérarchique des flux vidéo Moments de Zernike spatio-temporels Points SIFT Index. décimale : 004 Informatique Résumé : Optimization is a scientific discipline that is concerned with the extraction of optimalsolutions for a problem, among alternatives. Many challenging applications inbusiness, economics,and engineering can be formulated as optimization problems.However, they are often complexand difficult to solve by an exact method within areasonable amount of time.In this thesis we propose a novel approach called population based iterated greedy algorithm in order to efficiently explore and exploit the search space of one of the NP-hard combinatorial optimizationproblems namely the minimumn weigh vertex cover problem. It isa fundamental graph problem with many important real-life applicationssuchas, for example, in wireless communication, circuit design and network flows.An extensive experimental evaluation on a commonly used set of benchmarkinstances shows that our algorithm outperforms current state-of-the-art methods notonly in solution quality but also in computation time. Côte titre : DI/0010 En ligne : http://dspace.univ-setif.dz:8888/jspui/handle/123456789/35

Exemplaires (2)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
DI/0010	DI/0010-0011	Thèse	Bibliothéque des sciences	Anglais	Disponible Disponible
DI/0011	DI/0010-0011	Thèse	Bibliothéque des sciences	Anglais	Disponible Disponible

---

1 (1 - 1 / 1)