Skip to Main content Skip to Navigation
Theses

Traitement de l’imperfection des données temporelles saisies par l’utilisateur - Application aux logiciels destinés à des patients malades d’Alzheimer

Abstract : The need to deal with temporal data imperfection in the Semantic Web is increasing and, therefore, calls for a standard way of representing and reasoning about such data. This thesis presents a new vision aiming to provide solutions for different types of imperfections affecting temporal data, which we first define in a typology that we propose. Our thesis includes several contributions. The first one consists in defining a typology of temporal data imperfection in which we elaborate the various imperfections that can affect temporal data. We classify them into direct imperfections and others that are indirect. Then, we move to dealing with these imperfections in ontology. We propose an approach to deal with imprecision by representing and reasoning about dates and time clocks and associated qualitative relations in ontology with a crisp view. Therefore, we extend the 4D-Fluents approach for the representation and the Allen’s interval algebra for the reasoning. Then, we create an ontology based on our extensions called « CrispImpTimeOnto ». When dealing with uncertain temporal data in ontology and associated qualitative relations, we first propose an approach in a certain environment. We calculate certainty degrees using Bayesian networks. We extend the 4D-Fluents approach to represent the handled data and extending Allen’s interval algebra for the reasoning part. Second, we propose an approach base on possibility theory to calculate the necessity measures of the handled data and by extending the same approaches for representation and reasoning. Then, we conduct a small comparison between the two proposed approaches. We offer an ontology based on our extensions called “UncertTimeOnto”. Temporal data may, also, be affected by multiple types of imperfections at the same time. Thus, we present a contribution which consists in dealing with simultaneously uncertain and imprecise temporal data in ontology using possibility theory. We propose an ontology, named « UncertPossibTimeOnto », based on our extensions. We deal with the conflict of temporal data in ontology by using the evidential theory to calculate the belief masses of the handled data. We propose a semantic representation based on the obtained measures. We also extend Allen’s algebra to reason about conflicting temporal data. We propose an ontology, named « BeliefTimeOnto » based on our extensions. Based on the proposed ontologies in the previous chapters, we propose an ontology of temporal data imperfection called « TemporalOntoImperfection » that encompasses all the treated imperfections. Our contributions are implemented and evaluated within CAPTAIN MEMO, the memory prosthesis based on the PersonLink ontology, dedicated to Alzheimer’s patients. We integrate our temporal ontology « TemporalOntoImperfection» in the prosthesis and we evaluate our approaches dealing with the different types of imperfection. This assessment involves people with Alzheimer’s disease.
Document type :
Theses
Complete list of metadata

https://tel.archives-ouvertes.fr/tel-03504653
Contributor : Nassira Achich Connect in order to contact the contributor
Submitted on : Wednesday, December 29, 2021 - 3:09:15 PM
Last modification on : Thursday, January 6, 2022 - 9:33:11 AM

File

Rapport_Nassira_VF.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-03504653, version 1

Collections

Citation

Nassira Achich. Traitement de l’imperfection des données temporelles saisies par l’utilisateur - Application aux logiciels destinés à des patients malades d’Alzheimer. Informatique [cs]. HESAM Université; L’Université de Sfax, 2021. Français. ⟨tel-03504653⟩

Share

Metrics

Les métriques sont temporairement indisponibles