A CONNECTIONIST APPROACH FOR SIMILARITY ASSESSMENT IN CASE-BASED REASONING SYSTEMS

Case-Based Reasoning (CBR) systems support ill-structured decision mak ing. In ill-structured decision environments, decision makers (DMs) di ffer in their problem solving approaches. As a result, CBR systems wou ld be more useful if they were able to adapt to the idiosyncrasies of individual decision makers. Existing implementations of CBR systems ha ve been mainly symbolic, and symbolic CBR systems are unable to adapt to the preferences of decision makers (i.e., they are static). Retriev al of appropriate previous cases is critical to the success of a CBR s ystem. Widely used symbolic retrieval functions, such as nearest-neigh bor matching, assume independence of attributes and require specificat ion of their importance for matching. To ameliorate these deficiencies connectionist systems have been proposed. However, these systems are limited in their ability to adapt and grow, To overcome this limitatio n, we propose a distributed connectionist-symbolic architecture that a dapts to the preferences of a decision maker and that, additionally, a meliorates the limitations of symbolic matching, The proposed architec ture uses a supervised learning technique to acquire the matching know ledge. The architecture allows the growth of a case base without the i nvolvement of a knowledge engineer. Empirical investigation of the pro posed architecture in an ill-structured diagnostic decision environmen t demonstrated a superior retrieval performance when compared to the n earest-neighbor matching function. (C) 1997 Elsevier Science B.V.