Profile Adaptation in Adaptive Information Filtering: An Immune Inspired Approach

Abstract

Within the context of information filtering, learning and adaptation of user profiles is a challenging research area and is, in part, addressed by work in Adaptive Information Filtering (AIF). In order to be effective in a dynamic context, maintaining filtering performance, information filtering systems need to adapt to changes. We argue that artificial immune systems (AIS) exhibit the properties required by AIF, and have the potential to be exploited in the context of AIF. In this paper, we extract general features of immune systems and AIF, based on a principled meta-probe approach. We then propose an architecture for AIF incorporating ideas from AIS. Having such characteristics as adaptability, diversity and self-organised, we argue that AIS have suitable characteristics that are amenable to the task of AIF.