Modélisation et utilisation de ressources et services Web et indexation de données dans un contexte d’incertitude.

Asma Omri 1
1 SOC - Service Oriented Computing
LIRIS - Laboratoire d'InfoRmatique en Image et Systèmes d'information
Abstract : It is commonly accepted that data production has several years, a spectacular development due to the multiplication of new technologies such as social networks, new mobile devices, smart meters, sensors and cloud computing. In fact, this explosion of data should continue and even accelerate. Wondering how we should deal with this mass of which becomes more and more varied, complex and less structured, is so essential. DaaS (Data As A Service) can be defined as the supply, the management and provision of data presented in an immediately consumable format business users of organizations as a service. The returned data These services are generally characterized by uncertainty and heterogeneity. There are many approaches that process data according to the life cycle of the service Web based on 6 phases namely creation, selection, discovery, modeling, invocation and composition of services, in order to solve the volume problem data, its heterogeneity or its speed of evolution. On the other hand, there is very little approaches that address data quality and uncertainty processing in the Web. Naturally, in this thesis we have been interested in the question of Web in a context of distributed and heterogeneous systems. The main contribution to bring in this research work is to study the composition of services and / or web resources and indexing data in an uncertain context. As a first step, through the contributions of the literature, the theoretical framework on the specificities of the DaaS concept of uncertain service, is presented by adopting the possibilistic theory. The problem of web services composition and the impact of uncertainty, which can be associated with the exit of a service, on the processes of selection and composition of the services are explained. To do this, we proposed an approach Possibilist to model the uncertainty of data returned by uncertain services. Specifically, we have extended Web service description standards (for example, example, WSDL) to represent the uncertainty levels of the outputs. We also have extended the service invocation process to take into account the uncertainty of input data. This extension is based on the theory of possible worlds used in the possibilistic databases. We also put forward a set of composition operators, sensitive to uncertainty values, in order to orchestrate uncertain data services. In a second step, we studied the impact of uncertainty on the representation and manipulating web resources. We defined the resource concept Web uncertain and proposed mechanisms of composition of resources. To do this, a pattern of uncertainty description through the concept of uncertain web resource has been presented. It is based on a probabilistic model where each resource can have several possible representations, with a certain probability. Finally, and finally, we have proposed methods of documentary indexingBig data type data. Initially, we adopted a approach of syntactic indexing of uncertain data, then we followed a uncertain semantic indexing method. Finally, and to boost this approach, we proposed a hybrid method of indexing in an uncertain context.
Complete list of metadatas

Cited literature [144 references]  Display  Hide  Download
Contributor : Asma Omri <>
Submitted on : Monday, December 3, 2018 - 5:16:50 PM
Last modification on : Friday, December 21, 2018 - 1:36:30 PM
Long-term archiving on : Monday, March 4, 2019 - 2:59:11 PM


Files produced by the author(s)


  • HAL Id : tel-01943392, version 1


Asma Omri. Modélisation et utilisation de ressources et services Web et indexation de données dans un contexte d’incertitude.. Web. Université Claude Bernard Lyon 1; Université de Sousse, 2018. Français. ⟨tel-01943392⟩



Record views


Files downloads