Digestion of Knowledge in a KM System to Reveal Implicit Knowledge

Abstract

The motivation of this work is the problem of information overload in the ICT-based systems. We think network knowledge management systems have the most important characteristics of systems with this problem, but they are more scalable and controllable than others, so they might be used as a research experimental model. Our assumption is that there are several hidden aspects in the systems with information overload, which can be used to try to solve this problem. On the one hand, taking advantage of the excess energy of the active elements that are involved in the systems, such as users, services or applications and other entities related to them. One the other hand, using the properties of both the elements and the activities related to the systems affected by the problem, such as network, active entities, information and knowledge involved, or processes and interactions of these elements and activities. In applying this assumption to proposed simplified experimental model of knowledge management systems, we try to discover ways to reduce information overload in these systems, which could be applied in broader areas such as the Web. Our approach is based on a knowledge management system called KnowCat (KC) (Alaman & Cobos, 1999; Cobos, 2003; Cobos & Pifarre, 2008), which is a groupware system that facilitates the management of a knowledge repository by means of user community interaction through the Web. KC achieves a selection of the best documents without supervision by anyone, using information about users’ activity and users’ opinion about knowledge. KC knowledge repository is formed by documents and topics structured as a knowledge tree. Each KC instance is a KC Node and has a subject, a user community and a knowledge repository. Crystallization is the KC’s process of knowledge selection by its quality using information about users’ activity and their opinion about the knowledge items. For more information, please see the chapter about KC in this book. In order to support the assumption, a prototype has been developed on KnowCat, which is called Semantic KnowCat (SKC) (Moreno-Llorena, 2008; Moreno-Llorena & Alaman, 2005) incorporating ideas and techniques from different research areas that converge on the Semantic Web (Berners-Lee, 2000): Knowledge Management, Human Computer Interaction, Collaborative Work, and Data and Information Mining (Baeza & Ribeiro, 1999). This article shows how some of the techniques and ideas mentioned are integrated to implement an Analysis Module (AM) of SKC on a KC system. This module is in charge of processing explicit knowledge of the system in order to develop another latent and return it