烟支重量反馈控制系统相关参数的海量数据分析  

作　　者：孔祥昊 裘哲勇[2] 赵纪政 沈梓崴 李冠华 

机构地区：[1]杭州电子科技大学卓越学院,浙江杭州310018 [2]杭州电子科技大学理学院,浙江杭州310018

出　　处：《科技创新与应用》2018年第21期12-15,共4页Technology Innovation and Application

基　　金：浙江省新苗人才计划项目(编号:2017R407055)资助

摘　　要：文章采用统计学理论分析了卷烟过程中烟支重量反馈控制系统所监测的多项指标的海量数据。首先,对海量数据进行预处理,并通过假设检验得出数据服从正态分布。其次,计算各项指标间的灰关联度,得到各项指标间的关联度矩阵。最后,通过逐步回归分析得到各项指标间具有显著相关性的最优回归方程,并给出能够根据各监测指标的变化对烟支重量进行反馈调节的控制系统模型。In this paper, the statistical theory- is used to analyze the mass data of many indexes monitored by cigarette weight feedback control system in the course of cigarette production. First of all, the massive data are preprocessed, and the hypothetical test shows that the data obeys the normal distribution. Secondly, the grey con＇elation degree of each index is calculated, and the correlation degree matrix of each index is obtained. Finally, through stepwise regression analysis, the optimal regression equation with significant correlation among the indicators is obtained, and the control system model which can adjust the cigarette weight feedback according to the changes of each monitoring index is given.

关 键 词：海量数据 正态分布 灰关联度分析 逐步回归分析 

分 类 号：TS47[农业科学—烟草工业]