Print

 

Quarterly News report  4 – December 2010

File : Internet and Arabic Language

 

The information and communication technology (ICT) materialized especially through the Internet and combined with the widespread use of computers in all sectors have led to the proliferation of information sources distributed, involving considerable volumes of  heterogeneous information. Therefore, the development of automated tools for the effective access to such information is a must and is today a challenge to overcome.

 

This need has generated investigations and researches for the development of new mediators between information sources and users, including the information filtering systems. An information filtering system (SFI) is defined as process enabling to select and only spread information of interest to a user or group of users from a large volume of dynamic information. It allows targeting and delivering continuously relevant information that meets the needs of users.

 

 

The information filtering (IF) is neither a new concept nor a concept limited exclusively to electronic documents. Indeed, when we read any text on paper or other medium, the filtering process is involved. Thus, we are to filter parts of the huge body of information we have access to and that, whenever one wishes to gain some information.

 

The information filtering process is very similar to the information search process (IS), in the sense that they have the same goal which is to find information relevant to a user. They differ in the nature of the users’ information needs, and the flow of information. For IS, the need for information is ephemeral and the documentary background is stable. For IF, the need for information is stable and the document flow is dynamic.

 

For their role as mediators, information filtering system must have the methods and the knowledge to treat, assess, filter and extract relevant information for the user. It will:

Take into account the heterogeneity of information (structured, semi structured and non structured  documents)

Take into account the multilingual information available on the Internet.

Take into account the user's profile, in terms of interests on the spatiotemporal localization of his application.

 

Evaluate information according to certain quality criteria: the content (accuracy of the information, current information or not, single or not ...), on the responsible persons (institutional affiliation, author identification, ... ), the organization of information and accessibility (text or not, fast loading or not ...)

etc.

This file is intended to provide a general overview of the field of information filtering. It presents the information selection process likened to the filtering process. It tackles the fundamental issues facing a filtering system, describing its characteristics, the different forms of filtering, as well as different areas of application. It describes, in the end, some well-known public domain tools and some tools developed at the CERIST and some current research topics related to the field of filtering and access to information.

Nouali Omar

Research Director

Division of Theory and Engineering of Computer Systems