基于单层感知器的辨识方法及其应用

单层感知器神经网络模型是多层感知器神经网络———ＢＰ网络的基础，对单层感知器学习算法的改进是进行ＢＰ网络学习算法改进的基础。把带遗忘因子的递推最小二乘辩识算法的原理应用到单层感知器的学习算法中，提出了单层感知器的改进学习算法。这一改进算法克服了常规学习算法不适于在线学习的缺点。仿真实验的结果证实，基于改进学习算法的单层感知器完全可以满足线性系统在线辨识的要求。最后分析了这种改进算法的优点及其具有这些优点的原因

Indentification Method Based-on Single-Layer Perceptron and It s Application

Single-layer perceptron neural network is the base of multil-layer perceptron neural network-BP networks. The progress of sigle-layer perceptrons learning algorithm is the base of that of BP networks learning algorithm. An advanced learning algorithm of single-layer perceptron is given by applying the theory of recursive least square algorithm with a weight parameter into single-layer perceptrons learning algorithm. This advanced algorithm overcomes the disadvantage of the conventional learning algorithm that doesnt adapt to real-time learning. Experimental result demonstrates that single-layer perceptron based on advanced learning algorithm is completely meets the need of the linear systems real-time identification. Finally, this advanced algorithms advantages are analyzed.