基于Kalman滤波法的Madaline网络算法及其在部分未知背景干扰样品分析中的初步应用

Development of Madaline with Kalman Filter and its Application in Unknown Background Mixture

基于Kalman滤波法的Madaline网络算法的特点是训练与预测同时进行 ,在由已发现物质组成的未知背景干扰样品分析中具有较大的应用价值 ;可减少训练和预测的计算时间。

This paper introduce a novel method combining Kalman filter and Madaline network. Its good performance over the existed method is that it can predict the components involved in unknown background in the meantime of training procedure. It can save a lot of time in training procedure as well as give the corred results. A simulated mixture's analysis proved the above performance.