Article Preview
Top1. Introduction
Knowledge is an important asset of enterprises and many methodologies have been proposed to acquire and represent experiential knowledge, like CommonKADS (Shreiber, Wielinga, de Hoog, Akkermans, & Van de Velde, 1994) and MIKE (Angele & Studer, 1998).
These methodologies are devoted to manage structured knowledge through the definition of a well defined and formalized knowledge model. Unfortunately, there are a lot of fields and problems which cannot be described by standard and general methodologies, since their structures are so complex that it isn’t possible to create a model of the involved knowledge.
In this context, a very promising subject of investigation concerns the adoption of Case Based Reasoning (CBR) (Kolodner, 1993) methodology as a very suitable paradigm to deal with complex knowledge structures. It has been applied to many research areas (Muñoz-Avila, Gupta, Aha, & Nau, 2002; Watson, 2002b, 2002a) and it results to be the most natural approach for many research projects characterized by episodic knowledge, since it allows to find a solution to a new problem (i.e., the case) by the adaptation of solutions adopted in the past to solve the most similar problems to the current one.
The key aspects of a Case Based application are the structure of the case and the nature of the similarity among cases. The structure of the case must define what kind of attributes has to be adopted to describe problems, in order to allow the comparison among them. The nature of the similarity must define a criterion, that can be a function (i.e., similarity function) or something more complex like a similarity metric (Finnie & Sun, 2002), to compare cases according to their attributes, in order to determine what are the most similar past problems to the current one.
The definition of case structure and similarity function is rather simple when the examined domains are characterized by well defined knowledge: in such situations, the classical K–Nearest Neighbor based on the Euclidean Distance is sufficient to compare cases having a fixed and unchangeable structure. Unfortunately, there exist a lot of situations in which the structure of the case is not unique, since it depends on the context in which it is analyzed and more sophisticated similarity metrics than the K–Nearest Neighbor algorithm could be necessary to compare them.
The aim of the paper is to describe a conceptual and computational framework for the management of collective creativity in decision–making processes that cannot always be captured exploiting traditional methodologies for the development of knowledge–based systems. Such knowledge typically concerns informal groups of people working and living within organizations, called Communities of Practice (CoPs) (Wenger, 1998) as well as organizations with a low level of technicality, e.g., Small and Medium enterprises (SME).
Given the importance of SMEs in the global economy, especially in Europe, there has been a great deal of research in the Knowledge Management (KM) context, both from the theoretical and the practical standpoint to support SMEs in their day to day activities and to join SME networks. The main KM issues that have been recognized for SMEs are:
- •
Their significant technological gaps with respect to wider organizations. Several technological solutions have been proposed in KM literature to enhance networking and knowledge sharing within collaborative communities (see KNOW-CONSTRUCT project (Soares, Simões, Silva, & Madureira, 2006) as an example); and
- •
Their vulnerability in terms of loss of key personnel, as a consequence of their small size (Handzic, 2004). The limited dimensions of SMEs, that undoubtedly is a benefit from the agility perspective, may in fact cause the lack of a shared structured framework for company experiential knowledge collection and management.