All Title Author
Keywords Abstract


Keywords: 郑中义*(1964-),,博士,教授,博士生导师,

Full-Text   Cite this paper   Add to My Lib


为研究船舶领域与其影响因素的关系,首先, 对船舶领域随不同影响因素的变化进行统计分析;其次,分析船舶领域随其影响因素变化的本质,并利用小波变换对船舶领域样本数据的变化规律进行分析;最后,利用神经网络模型对其间关系进行建模.建模时, 选取小波变换后的2/3数据进行训练建模,选取剩余1/3数据对模型模拟的能力及鲁棒性进行验证,获得每个影响因素值下的船舶领域平均值及基于神经网络的船舶领域与其影响因素关系模型.结果表明,该模型具有很好的性能,其拟合度R为0.92242,平均误差为49.8970 m,且具有较好的鲁棒性.
In order to study the relationship between the ship domain and its influence factors, firstly, the changes of ship domain with different factors were statistically analyzed. Secondly, the changes nature of the ship domain with its influence factors were analyzed, and by using wavelet transform, the change rule of the sample data of ship field was studied. Finally, the modeling of correlation was conducted by using artificial neural network (ANN). In modeling process, the 2/3 data of wavelet transform was chosen for the training, and the simulating ability and robustness of the model were verified by the rest of 1/3 data. The average value of ship domain under every influence factors and the model of relationship between ship domain and its influence factors based on the artificial neural network were obtained. Results show that the model of ship domain has a good performance with a fitting degree of R 0.92242, and an average error 49.8970 m, and a better robust.


comments powered by Disqus