La recherche d’informations par accès aux contenus sémantiques: vers une nouvelle classe de Systèmes de Recherches d’Informations et de moteurs de recherche (aspects linguistiques et stratégiques)
Current Information Retrieval Systems (IRS), which operates primarily on terms (by measuring their presence and co-presence or their absence) and without using semantic relations, increases the number of documents thus causing an unquestionable noise that is difficult to control by the users. How to discover rare or hidden document, which contains genuine information? How to select relevant information for precise goals? To introduce more semantics in information retrieval paradigm for textual documents we propose a new approach based on a new linguistic and computational technique of Exploration of the Context, developed at the LaLIC laboratory, which is more complex than the identification of regular expressions. It comes to modeling and carrying out a new system generation for Information Retrieval Systems (IRS) which is founded on automatic annotations of texts. We present in this paper, the interaction with an indexing process throw building a prototype which results from a coupling of a two engines: (i) an annotation system EXCOM and (ii) a second indexing system MOCXE. The consequences of these IRS are economically fully serviced. The various fields of survey (strategic, economic, innovation...) have obvious needs for powerful data-processing tools which would enable them to detect rare information rather than to consolidate only known information of all the specialists. The innovation indeed supposes the detection of signals which indicate, for example, the emergence of new concepts, new methods, new molecules, new techniques, the unknown relationships for companies, and the nomination of such new leader, new potential markets, new requests for users, a beginning of epidemic or a crisis.
Lista citărilor/referințelor nu cuprinde decît texte prezente în baza de date, nefiind deci exhaustivă. Pentru trimiterea de texte, semnalarea oricăror greșeli, și eventualul refuz ca „Diacronia” să facă publice textele, vă rugăm să folosiți adresa de email [Please enable javascript to view.].
pdf 
html 
