线性EIV模型参数的LASSO估计方法

来源 :山东理工大学学报(自然科学版) | 被引量 : 0次 | 上传用户:x1026221496
下载到本地 , 更方便阅读
声明 : 本文档内容版权归属内容提供方 , 如果您对本文有版权争议 , 可与客服联系进行内容授权或下架
论文部分内容阅读
基于结构风险最小化原则,提出了可以实现高维数据降维的线性EIV模型参数的LASSO估计(LE)方法,并给出了其数值解的迭代算法.为说明LE方法的有效性,通过实证与WTLS、LS两种方法进行了对比分析.结果表明,LE方法能够明显提高预测精度,具有更强的泛化能力,同时可以实现变量选择,达到高维数据降维的目的.
其他文献
Coanda jet flap is an effective flow control technique,which offers pressurized high streamwise velocity to eliminate the boundary layer flow separation and increase the aerodynamic loading of compressor blades.Traditionally,there is only single-jet flap
The supersonic multi-hole probe is an essential test tool for wind tunnel experiments,which is necessary to develop the basic research of improving the measurement accuracy and expanding the application of the probes.This paper theoretically derived a gas
Since the transition from rotating stall to surge in a transonic compressor at high speed is very quick,quite often there is no time to take measures to prevent the surge.Therefore,it is desired to find any rotating stall precursors,of which the occurrenc
The S-shape inlet has been widely used in advanced military aircraft due to their advantage of reducing radar signature.However,the curvature of the inlet usually causes different kinds of intake distortion at the aerodynamic interface plane (AIP).Among t
A coupling configuration of double-wall cooling and exterior surface thermal barrier coating (TBC) is one of the most promising thermal protection methods of hot components of modem gas turbine.The combined influences of coating thickness,impingement layo
针对船舶环境亮度的不同,通过摄像机获取船舱环境图片信息,将HIS模型与Retinex理论结合获取图像照度模型,从而自适应控制船舱LED灯照度.将RGB空间中的彩色图像映射到HIS空间,根据瑞利散射特性获取不同颜色通道权重,得到改进后HIS亮度分量模型;应用Retinex理论将亮度分量与高斯中心环绕函数做卷积运算,最终得到图像照度估计.实验结果表明,与传统算法相比,本文算法减少了颜色分量干扰,能获取到更为精确的照度信息.
研究了现有的传统GM(1,1)(简称GM(1,1))、基于缓冲算子和时间响应函数优化的GM(1,1)、首输入GM(1,1)(简称FGM(1,1))、变权缓冲GM(1,1)(简称VWGM),针对它们存在的问题,提出了一种新的预测模型——混合GM(1,1)(Hybrid GM(1,1),简称HGM(1,1)).该模型引入了粒子群优化算法、变权缓冲算子以及FMG(1,1)模型,并以拟合值与实际值的灰色关联度最大为目标进行参数优选.结果显示,改进后的模型增强了适应性,能够充分利用原始数据信息,提高模型的预测精度,
为提高还原型谷胱甘肽的稳定性和生物利用度,以卵磷脂和胆固醇为膜材料,采用薄膜-超声法制备还原型谷胱甘肽纳米脂质体(GLip).以包埋率为指标,在单因素实验的基础上,利用响应面法优化GLip制备的工艺参数.实验得出的最优制备条件如下:超声时间为10 min,卵磷脂添加量为232 mg,卵磷脂与胆固醇的质量比为4∶1,谷胱甘肽添加量为75 mg,吐温-80的添加量为94 mg;在此条件下得到的最大包埋率为(65.85±0.54)%.优化后的高包埋率还原型谷胱甘肽纳米脂质体在食品工业和医药等领域具有广阔的应用前
通过设计一个自适应控制器和参数更新法则,实现两个分数阶不确定耦合递归动态神经网络系统(fractional-order uncertain coupling recurrent dynamical neural networks,FUCRDNNs)的同步和参数识别;构造辅助函数证明了所提出方法的局部渐近稳定性,并得到了两个相关推论;最后通过两个数值算例表明了该方法的有效性.
The plasma actuation (PA) effect on the snow falling flow was investigated using a plasma electrode with weather resistant design and the natural snow wind facility of the Hokkaido University of Science.NACA0015 test blade with chord length c of 300 mm wa