Lazy Propagation in Case Retrieval Nets.
Mario Lenz, Hans-Dieter Burkhard · 1996
. An efficient retrieval of a relatively small number of relevant cases from a huge case base is a crucial subtask of Case-Based Reasoning (CBR). In this article, we investigate Case Retrieval Nets, a memory model for this task which is applicable in any domain with attribute-based case representations. The main idea is to apply a restricted spreading activation process in the case memory in order to retrieve the cases most similar to a posed query case. 1 INTRODUCTION An efficient retrieval of relevant cases is a crucial subtask of CaseBased Reasoning (CBR). In recent years, a number of techniques have been proposed, such as indexing techniques ([11]), kd--trees ([13, 14]), the heuristic FISH--AND--SINK approach ([12]), or the CRASH memory model ([3]). In order to increase efficiency, these techniques either limit the flexibility of case retrieval or require a large amount of structured world knowledge. In [4, 5, 9] we developed Case Retrieval Nets (CRN) as a memory model for an ef...