摘要
分析了我国海监船和渔政船发展现状及发展趋势。整理分析相关的船型数据资料,利用逐步回归的数学方法,分别基于Excel回归分析、MATLAB的BP神经网络工具箱和RBF神经网络工具箱,建立了海监船和渔政船主尺度的数学模型,并对3种数学统计模型进行了实船验证和误差分析,结果显示BP神经网络和RBF神经网络模型误差较小。数学模型的建立有利于分析和掌握海监船与渔政船主尺度变化的规律,为报价设计和初步设计提供了科学依据。
The present situation and tendency of the fishery patrols of our country are analyzed in this paper.Based on the statistics of the fishery patrols principal dimensions data,the mathematical models of the principal dimensions are established respectively using the stepwise regression method,the BP Neural Networks Toolbox and the RBF Neural Networks Toolbox of MATLAB.By analyzing the results of the comparative results,the result of RBF model is proved to be the best.The results indicate that the models are reliable and available by verifying the models on fishery patrols.The variation of the principal dimensions of the fishery patrols is helpful in guiding the quote design and the compact design.
出处
《舰船科学技术》
北大核心
2012年第7期49-54,共6页
Ship Science and Technology
基金
国家公益性行业科研专项:渔业节能关键技术研究与重大装备开发(201003024)
关键词
海监船
渔政船
主尺度
数学模型
回归分析
神经网络
fishery patrol
principal dimension
mathematical model
regressive analysis
neural networks
