Supporting Knowledge Management using a Nomadic Service for Artifact Recommendation
Pablo Gomes Ludermir · 2005
Knowledge Management (KM) can be defined as the effective strategies to get the right piece of knowledge to the right person in the right time. Having the main purpose of providing users with information items of their interest, recommender systems seem to be quite valuable for organizational knowledge management environments. Early KM attempts relied on centralized knowledge bases and have shown a number of drawbacks. For instance, centralization forces users to agree in a common classification for the whole organizational knowledge. This disrespects the personal and distributed nature of knowledge. One approach to solve this problem is the usage of agents to represent actors on KM scenarios, showing effectiveness due to the autonomy, proactive behavior and sociability of agents. Current mobile devices and wireless network technologies are able to offer information about their context. With that we can think of new kinds of reconfigurable services that take into account the user’s context, i.e. nomadic services. Nomadic mobile service provisioning is a promising new paradigm for service provisioning, where individuals can act as a service provider, offering services on an ad-hoc basis to other users. This paradigm opens up new possibilities for knowledge management systems and services. This Master Thesis describes the design and implementation of KARe (Knowledgeable Agent for Recommendations), a multi-agent recommender system that supports nomadic users sharing knowledge in a peer-to-peer environment with the support of a nomadic service. Central to this work is the assumption that social interaction is essential for the creation and dissemination of new knowledge. Supporting social interaction, KARe allows users to share knowledge through questions and answers. Furthermore, we assume that nearby users are more suitable for answering his questions in some scenarios and we use this information for choosing the answering partners during the recommendation quest process.