Personnalisation et enrichissement des méthodes d’accès aux données

Grégory Smits 1
1 SHAMAN - Symbolic and Human-centric view of dAta MANagement
IRISA-D7 - GESTION DES DONNÉES ET DE LA CONNAISSANCE
Abstract : The translation of data into knowledge is a crucial task at the heart of many professional activities. Two main strategies may be envisaged to perform this translation : by querying a database management system or by using data mining techniques. These two approaches have been so far studied independently by two distinct communities, namely the database community and the data mining one. The works described in this document, whose aim is to synthetize the research results obtained during the last eight years passed in the IRISA laboratory, mainly belong to the database area. However, considering the growing importance of unstructured data, my last contributions are at the intersection of data mining and databases. The common thread in this document is the enrichment of the methods used to access data. Data access is considered as a three steps process : 1) the expression of an information need, 2) the efficient retrieval of data satisfying the considered information need, and 3) the restitution of the query results to the user. The singular aspect of the data processing chain described in this document relies on the leading role given to the user at each step of the process defined to translate data into knowledge. The first part of the document is dedicated to the enrichment of some methods used to access data. My contributions on that point are twofold. The first one aims at making querying interfaces more flexible and at increasing their expressivity by letting users access data using their own vocabulary composed of linguistic terms. The second approach consists in helping users, with cooperative strategies or intuitive query interfaces, translate their information needs into queries. As commercial database systems do not provide flexible querying functionalities, the second part of the document describes my contributions on the evaluation of selection statements involving conditions based on the satisfaction of subjective linguistic terms. Through these last works, I have shown that a compromise may be found between flexibility and efficiency when querying data. An intelligent data management system should also assist users during the analysis of the results of their queries. Cooperative answering strategies aim at helping users understand the content of a result set and also aim at enriching it with indirect answers and complementary knowledge. The third part of the document details several cooperative answering strategies that ease the translation of query results into knowledge. The theoretical framework that links the different parts of the data processing chain presented in this document is soft computing. In this sense, an underlying objective of this document is also to show that the theories and techniques of soft computing bring pragmatic and innovative solutions to answer the crucial issue of data management. A positive conclusion and perspectives for future research directions are given at the end of this document about the role the soft computing community can play by promoting the idea of representing, computing and reasoning about data with words.
Document type :
Habilitation à diriger des recherches
Complete list of metadatas

Cited literature [222 references]  Display  Hide  Download

https://hal.inria.fr/tel-01739707
Contributor : Grégory Smits <>
Submitted on : Wednesday, March 21, 2018 - 12:01:34 PM
Last modification on : Friday, January 11, 2019 - 2:28:12 PM
Long-term archiving on : Thursday, September 13, 2018 - 7:13:43 AM

File

main.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01739707, version 1

Citation

Grégory Smits. Personnalisation et enrichissement des méthodes d’accès aux données. Base de données [cs.DB]. Université Rennes 1, 2018. ⟨tel-01739707⟩

Share

Metrics

Record views

476

Files downloads

340