Knowledge Acquisition for Business Intelligent Systems

Knowledge acquisition is a frequent bottleneck in business intelligent systems, and knowledge extraction from trained artificial neural networks (ANNs) provides an excellent way for explaining the functioning of a business connectionist system. This is important for ANNs to gain a wider degree of acceptance in problem domain applications like classification, diagnosis, continuous function approximation, time series prediction, and data mining. All these typical applications could be implemented into hybrid business intelligent systems. This paper presents the new applied research techniques for extracting business knowledge from trained ANNs, and it is organized into four sections that include an introduction, the state of the art, knowledge extraction and representation by rules, and conclusions