Neural Network Support in a Hybrid Case-Based Forecasting System

The progress to date is described in an ongoing project in which the aim is to investigate the combination of case-based reasoning and artificial neural networks as a strategy for cooperative problem solving. The paper describes a successful application in which a Radial Basis Function artificial neural network is used for the adaptation of cases, during the reuse phase of the CBR life cycle. The approach is being applied to the problem of real-time oceanographic forecasting and the results obtained so far are presented.