基于域理论的自适应谐振算法FTART采用了独特的解决样本间冲突和动态扩大分类区域的方法,学习速度快、精度高,取得了很好的实用效果。但由于其学习到的知识是通过连接权隐式表示的,可理解性差,推理过程难以解释,使其进一步发展受到了很大限制。本文从功能性角度出发,提出了一种从FTART网络中抽取if-then规则的方法。实验结果表明,抽取出的if-then规则可理解性好、预测精度高,可以很好地描述FTART网络的性能。 The adaptive resonance algorithm based on domain theory FTART uses a unique method to solve the conflict between samples and dynamically expand the classification area. It has fast learning speed and high precision, and has achieved good practical results. However, due to the fact that the knowledge it learns is implicitly expressed through the connection right, its comprehensibility is poor and the reasoning process is difficult to explain, which limits its further development. This paper proposes a method of extracting if-then rule from FTART network from the perspective of functionality. The experimental results show that the extracted if-then rule has good comprehensibility and high prediction accuracy, and can well describe the performance of FTART network.