江西农业学报2023，Vol.35Issue(12)：47-52,6.DOI:10.19386/j.cnki.jxnyxb.2023.12.007
基于残差网络综合赋权法的卷烟物理质量评价方法研究
Evaluation Method for Cigarette Physical Quality Based on Residual Network Comprehensive Weighting
摘要
Abstract
In response to the low level of physics quality evaluation in the production and processing of finished cigarettes,as well as the lack of multi-index decision-making and weighting,a comprehensive weighting method based on residual networks is proposed for the analysis and evaluation of physical quality indicators of cigarettes.By introducing the Lambert distance,the conflicting weights of indicators evaluated by subjective and objective weighting methods were measured.The measured values of cigarette physical quality and expert scoring results were used as sample data for network training.The weights of cigarette indicators were iteratively trained through residual networks to ensure the objective authenticity of cigarette indicator weighting,retaining the emphasis of expert experience on significant indicators.Through experimental verification,this algorithm is more in line with the actual production situation in assigning weights to cigarette indicators,and can effectively assign reasonable weights to various physical quality indicators of cigarettes,thereby improving the quality of finished cigarettes.关键词
卷烟物理质量评价/指标赋权/残差神经网络/冲突度量Key words
Evaluation of cigarette physical quality/Indicator weighting/Residual neural network/Conflict measurement分类
轻工纺织引用本文复制引用
许冰洋,胡荣辉,刘赐德,钟智敏,许赟赟,廖英平,黄春杨,杨俊杰..基于残差网络综合赋权法的卷烟物理质量评价方法研究[J].江西农业学报,2023,35(12):47-52,6.基金项目
江西中烟工业有限责任公司科技项目(2020-09). （2020-09）
