Principes de conception de systèmes hypertextes pour la fouille de données bibliographiques multibases

Abstract : Information is essential in scientific and technical research and watch. The significant quantity of currently available data in a domain requires to implement adapted tools to exploit them. The goal of this research is to provide an environment in which the data of a domain (bibliographical references and Web) can be exploited for bibliographical search or domain analysis needs. In this framework, a general approach to build an hypertextual datamining system on bibliographical data is proposed. The use of hypextext capabilities favorizes an explorative access to data. Functionalities of datamining (statistical information, classifications, rules extraction) may be available to analyze data more precisely. The principal idea of this thesis is that datamining and information retrieval are two complementary approaches to access and analyse data : datamining allows to guide the information retrieval by using the knowledge extracted from the data. Conversely, information retrieval allows to guide the datami- ning pro cess by taking into account the extracted knowledge. The datamining process also favours the information access on the Web. Concretely, the knowledge extracted from bibliographical data provides a help for query formulation and improve the answer's precision of web search engine Building such a system requires the exploitation of different technics, Le. datamining, information retrieval and database management. From a technical point of view, the tools of these fields are combined thanks to a modular approach exploiting XML for the representation and the exchange of the data, and a data flow processing.
Document type :
Theses
File URL :
http://docnum.univ-lorraine.fr/prive/SCD_T_2001_0008_NAUER.pdf
Complete list of metadatas

https://hal.univ-lorraine.fr/tel-01746426
Contributor : Thèses Ul <>
Submitted on : Thursday, March 29, 2018 - 10:37:31 AM
Last modification on : Friday, May 24, 2019 - 10:58:08 AM

Identifiers

  • HAL Id : tel-01746426, version 1

Collections

Citation

Emmanuel Nauer. Principes de conception de systèmes hypertextes pour la fouille de données bibliographiques multibases. Autre [cs.OH]. Université Henri Poincaré - Nancy 1, 2001. Français. ⟨NNT : 2001NAN10008⟩. ⟨tel-01746426⟩

Share

Metrics

Record views

24