Project: Orientation in Complex Information Spaces as an Emergent Behavior of Autonomous Information Agents (IRAIA)

Orientation in Complex Information Spaces as an Emergent Behavior of Autonomous Information Agents (IRAIA)

Contact:

J. Lindlbauer

Project team:		Deutsches Institut f�r Wirtschaftsforschung (DIW), Berlin, coordination and database design Foundation for Research and Technology (Forth), Greece, similarity of time series Universit� Paul Sabatier (UPS), Toulouse, text analysis Queens University (QUB), Belfast, information visualisation Reticular Systems (Reticular), USA, internet agents Gesellschaft f�r iterative Softweretechnologien mbH (iteratec), Munich, marketing Ifo Institut f�r Wirtschaftsforschung (ifo), Munich, user requirements LemonLabs GmbH (LemonLabs), Munich, internetting software
Client:		European Commission / Information Society/ IST Programme (IST-1999-10602), shared cost project
Project period:		March 2000 to Feb. 2002

Information retrieval system for the Internet, focused on a highly interactive environment to assist the users by presenting the information space by a semantic coordinate system that provides them with a clear orientation during their retrieval process. Multilinguality is a prerequisite.

Methodology

Entry: Visualisation of the information contents of a data base. As an example a database for economic information was chosen. The visualisation is achieved by describing the information by their dimensions (region, sectors, themes, authors etc.) and letting the user choose among those (e.g. France/ mechanical engineering/ production, Ifo) or through a visualisation of the density of the information (Kohonen map, QUB).
Finding related information: The search keywords trigger an automatic cross language indexation and a term distribution analysis of the documents in the database. The information is then chosen using a ranked retrieval list (UPS). The text shows up on the users screen together with a list of the most important terms related to the text (all dimensions referenced within the text). Clicking on a term e.g. "mechanical engineering" opens the relevant structure (in the example "sectors") the user marks another term e.g. "textiles" and thus starts a new retrieval for related texts together with figures for the sector textiles (DIW). Starting from a time series, the user can also search for time series with a certain similarity (e.g. indicators; Forth).
Repeated search: The system gives the possibility to store the profiles of the searches. Thus the user can repeat the retrieval by hand or automatically. He can also send a query through the internet by "agents" (Reticular).
User validation: On different stages of the project the results are being shown to expert as well as casual users, for an evaluation and new ideas (Ifo).

Data

The system contains example information from Ifo (Business Survey figures and schnelldienst texts), DIW (National Accounts figures, Economic Bulletin texts) and Eurostat (LINDA database figures).

Results

concluded

Department conducting the study

Ifo Project

Orientation in Complex Information Spaces as an Emergent Behavior of Autonomous Information Agents (IRAIA)

Contents

Methodology

Data

Results