Information management in global environments: swarm intelligence in multilingual economic document repositories

Abstract

The information is a strategic resource of first order for organizations, so it is essential to have methodologies and tools that allow them to properly manage information and extract knowledge from it. Organizations also need knowledge generation strategies using unstructured textual information from different sources and in different languages. This paper presents two bio-inspired approaches to clustering multilingual document collections in a particular field (economics and business). This problem is quite significant and necessary to organize the huge volume of information managed within organisations in a global context characterised by the intensive use of Information and Communication Technologies. The proposed clustering algorithms take inspiration from the behaviour of real ant colonies and can be applied to identify groups of related multilingual documents in the field of economics and business. In order to obtain a language independent vector representation, several linguistic resources and tools are used. The performance of the algorithms is analysed using a corpus of 250 documents in Spanish and English from different functional areas of the enterprise, and experimental results are presented. The results demonstrate the usefulness and effectiveness of the algorithms as clustering technique.