论文部分内容阅读
利用凸集投影理论深入研究了函数链神经网络（ＦＬＮＮ）的学习问题，给出了相应的学习算法，并对算法的松弛形式进行了详尽的分析。由于采用了投影技术，网络的学习速度要比δ规则快得多，本文最后给出的仿真结果验证了该算法的稳定性和有效性。
The convex learning set theory is used to deeply study the learning problems of function chain neural network (FLNN). Corresponding learning algorithms are given and the loose form of the algorithm is analyzed in detail. As a result of the projection technology, the learning speed of the network is much faster than the δ-rule. The simulation results at the end of this paper verify the stability and effectiveness of the algorithm.