基于投影寻踪学习的大型船舶运动极短期预报

The Projection Pursuit Learning to the Extreme ShortPrediction of Big Ship Motion

本文结合投影寻踪和神经网络的优点 ,建立了应用于大型船舶运动的极短期预报的多维投影寻踪学习网络(PPLN)结构及算法 ,并将该算法所取得的预报结果与自回归预报法和周期图预报法的结果进行比较 ,预报结果说明了该算法的可行性。

This paper establishes a new projection pursuit learning networks to the extreme short prediction of big ship motion, which is a combined project pursuit learning and neural networks characteristic.In this paper we also give the algorithm. The prediction results are compared with auto-regress and periodical digram,and real example indicates that the algorithm is feasible.