论文部分内容阅读
本文提出了一种基于域理论的自适应谐振神经网络算法ＦＴＡＲＴ２，该算法将自适应谐振理论和域理论的优点有机结合，不需人为设置隐层神经元，学习速度快、精度高．此外，本文还提出了一种从ＦＴＡＲＴ２网络中抽取符号规则的方法，实验结果表明，使用该方法抽取出的符号规则可理解性好、预测精度高，可以很好地描述ＦＴＡＲＴ２网络的性能．
In this paper, an adaptive resonance neural network algorithm FTART2 based on domain theory is proposed, which combines the advantages of adaptive resonance theory and domain theory. It does not need to set the neurons of hidden layer artificially, so learning speed is fast and the precision is high. In addition, this paper also presents a method of extracting symbol rules from FTART2 network. The experimental results show that the symbol rules extracted using this method are understandable and have high prediction accuracy and can well describe the performance of FTART2 networks.