三次型火控回归解算模型变量筛选分析

Variable Selection of Thrice Fire Control Regress Calculation Model

在建立火控射击诸元回归解算模型时,由于影响火控射击诸元的因素较多,并且各因素之间存在交互作用。需利用逐步回归法对回归模型中的变量进行筛选,首先确定影响射击诸元的因素指标集,并划分因素水平。利用正交试验进行航路设计,并将数据装定到火控系统中,获得静态解算数据。利用逐步回归法在403个变量中选择14个变量通过多元线性回归得到火控回归模型。对火控模型进行偏F检验,对残差进行独立性和正态性检验。结果表明,通过逐步回归法筛选得到的模型变量建立的回归方程,回归效果较好。

When building a fire control shoot element regress calculation model, because many factors may influence fire control shoot elements, and each factor may interact on each other, the step by step regress is used to analyse the selected variables in the regress model. Firstly defining the factor collection that influences the shoot elements. Using the orthogonal test to carry out the design in route, then loading the data to the fire control system, and gaining static count data. Using the step by step regress to select 14 variables in 403 variables, then geting a fire control model by using the multianalysis linear regress. The fire control model has passed the partial F inspectation, and independence and normal form inspections are carried out for the residual error. The result indicates that the regress equation established by the step by step regress law has better regress effect.