Uncertainty management in connectionist expert system
Abstract
A connectionist expert system is an expert system whose knowledge base is generated from training examples using an artificial neural network learning technique. Gallant [13] developed a model for a connectionist expert system in which a variable is represented by a node and accepts two values, true or false. This study adopts two approaches to help manage uncertainty in Gallant's model. The first approach is called the random cell method while the second one is the stairstep method.
Collections
- OSU Theses [15752]