Multi-dimensional Information Ordering to Support Decision-making Processes

Massive amounts of textual and digital data are<br> created daily from business or public activities. The organisation,<br> mining and summarization of such a rich and large<br> information source is required to capture the essential and<br> critical knowledge it contains. Such a mining is of strategic<br> importance in many domains including innovation (eg to mine<br> technological reviews and scientific literature) and electronic<br> commerce (eg to mine customer reviews).<br> Information content generally bears several important aspects,<br> mapped onto visualisation dimensions, whose number<br> needs to be reduced to enable relevant interactive exploration.<br> In this paper, we propose a novel strategy to mine and organise<br> document sets, in order to present them in a consistent manner<br> and to highlight interesting and relevant information patterns<br> they contain.<br> We base our method on the formulation of a global optimisation<br> problem solved by using the Traveling Salesman Problem<br> (TSP) approach. We show how this compact formulation opens<br> interesting possibilities for the mining of document collections<br> mapped onto multi-dimensional information sets. We discuss<br> the issue of scalability and show that associated scalable<br> solutions exist. We demonstrate the effectiveness of our method<br> over several types of documents, embedded into real business<br> cases.