Development of Madaline with Kalman Filter and its Application in Unknown Background Mixture
Yu Chuan Wu · Jisuanji yu yingyong huaxue · 2000
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.