更为广义的线性回归模型及其气象应用

发布时间：2018-01-13 02:06

本文关键词：更为广义的线性回归模型及其气象应用　出处：《解放军理工大学学报(自然科学版)》2017年02期 　论文类型：期刊论文

【摘要】：为了解决传统的线性回归模型不具备全域分析能力以及表达能力受到模型维数限制的问题,基于要素场和矩阵的概念,提出了基于场的更为广义的线性回归模型——全回归模型。利用大规模要素场分布资料构建全回归模型,模型基本方程采用矩阵形式,方程的基本元素为要素场,相较于传统回归模型,该模型涵盖要素场整体信息。与传统回归方案进行对比分析,结果表明:该模型具有全局表达能力,在拟合过程中采用最小二乘方案可以得到全局最优结果;不同于传统回归模型的单调性特征,该模型自变量局部扰动对因变量的影响有利于分析相关过程;对于传统统计预报而言,基于矩阵分析理论及该模型全局表达能力,可以将该模型延伸应用至单站回归预报。
[Abstract]:In order to solve the problem that the traditional linear regression model does not have the global analysis ability and the expression ability is limited by the dimension of the model, the concept of element field and matrix is based on. A more generalized linear regression model-full regression model based on field is proposed. The full regression model is constructed by using the distribution data of large-scale factor field. The basic equation of the model is matrix form. The basic element of the equation is the element field. Compared with the traditional regression model, the model covers the whole information of the element field, and compared with the traditional regression scheme, the results show that the model has the global expression ability. In the process of fitting, the global optimal results can be obtained by using the least square method. Different from the monotonicity of the traditional regression model, the influence of the local disturbance of the independent variable on the dependent variable is beneficial to the analysis of the correlation process. For the traditional statistical forecasting, based on the matrix analysis theory and the global expression ability of the model, the model can be extended to the single-station regression prediction.
【作者单位】：解放军理工大学气象海洋学院;海军大连舰艇学院;解放军31110部队;解放军75839部队;
【基金】：中国科学院可再生能源重点实验室开放基金资助项目(Y707k31001) 中国科协高端科技创新智库青年项目(DXB-ZKQN-2016-019) 山东省自然科学基金资助项目(ZR2016DL09)
【分类号】：P731;P456
【正文快照】： 3.解放军31110部队,江苏南京,210016;4.解放军75839部队,广东广州,510510)气象海洋学的研究与预测方案主要依据动力学方法,相对于该方法,统计方法在大规模要素场的分析、预测中应用相对较少。对于广大基层气象海洋从业技术人员而言,统计方法在计算效率和应用层面上较数值模式

【相似文献】