Generalising Symbolic Knowledge in Online Classification and Prediction Richard Dazeley and Byeong-Ho Kang School of Information Technology and Mathematical Sciences, University of Ballarat, Ballarat, Victoria 7353, Australia. School of Computing and Information Systems, University of Tasmania, Hobart, Tasmania, 7001. r.dazeley@ballarat.edu.au, bhkang@utas.edu.au Abstract. Increasingly, researchers and developers of knowledge based systems (KBS) have been incorporating the notion of context. For instance, Repertory Grids, Formal Concept Analysis (FCA) and Ripple-Down Rules (RDR) all integrate either implicit or explicit contextual information. However, these methodologies treat context as a static entity, neglecting many connectionists’ work in learning hidden and dynamic contexts, which aid their ability to generalize. This paper presents a method that models hidden context within a symbolic domain in order to achieve a level of generalisation. The method developed builds on the already established Multiple Classification Ripple- Down Rules (MCRDR) approach and is referred to as Rated MCRDR (RM). RM retains a symbolic core, while using a connection based approach to learn a deeper understanding of the captured knowledge. This method is applied to a number of classification and prediction environments and results indicate that the method can learn the information that experts have difficulty providing. Keywords. hidden context, knowledge based systems, knowledge representation, ripple-down rules, situation cognition 1 Introduction Traditionally, knowledge based approaches have been based on the physical symbol hypothesis [1] which is built around the idea that knowledge is a substance that exists. However, after numerous failed systems some researchers have revised these concepts of knowledge and moved towards a situation-cognition (SC) based view. The SC view revolves around the premise that knowledge is generated at the time of its use. This implies that the existence of knowledge is based on the context of a given situation [2, 3]. A few methodologies, such as Formal Concept Analysis (FCA) [4], Repertory Grids [5] and Ripple-Down Rules (RDR) [6], have adopted a weak SC position by including contextual information. These approaches either incorporated