Scholarly record

EVALUATION OF INFORMATION RETRIEVAL SYSTEMS: CAN WE FIND MULTILINGUAL INFORMATION FROM DATA WAREHOUSES (WEB-CATALOGS AND WEB-DATA)?

O. Larouk, J. Dichy

Abstract

Access to digital data through web servers and databases is facilitated by search tools. Following a user query, the search engine has a long list of bibliographic references and web pages. The efficient selection of relevant documents is very difficult given the low accuracy in the list. Typically, the user visits the first page referenced in the list, but it does not consult the fiftieth. In this article, we describe the problem of the queries language when we look for information in an automatic system (search engine, database, Q/A systems, OPAC (Online Public Access Catalog), etc.). In most cases, we note today that research on the Internet is most often limited to the language of the user while the linguistic dimension opens an additional opportunity to have a wider application, that relevant should be incorporated in the process of information research if you want more information hidden multilingual. Multilingual Information Retrieval provides access to documents whose language is different from that in which the request is made. We propose to processing natural language using the contribution of reformulate the queries for computer-human dialogue. The reformulation of questions in natural language using a multilingual predicates (simple or complex), but the new systems of question/answers are a very open in the web content. Our experiments show that our combinatorial methods based on the coupling of question reconstructions with logical operators (BOOLE) and the metadata facets considerably reduce documentary noise and restrict irrelevant information.

Publication Impact Profile

No metrics available.

PLUMX see details

DOI resolver: doi.org/10.5593/sgemsocial2017/41/S17.041

Publication details

Back to publication list