基于BP神经网络的卷烟销售违规预测研究
发布时间：2019-07-02 14:44
【摘要】：烟草行业是高利润行业，卷烟打假是烟草专卖局的一个主要职责，由于信息技术的发展，现阶段如何提高卷烟打假的效率和准确度，实现精准打假，是一个需要认真探讨的问题。 论文将BP神经网络技术应用于烟草专卖检查工作中，探索卷烟销售违规预测模型，具有重要的理论与实践意义。通过对卷烟零售户的销售行为和销售心理进行定性分析，找出影响零售户违规行为的相关因素。针对某市烟草公司收集相关数据，对数据进行了预处理，对大量的历史销售记录和与之相关的数据进行定量分析，建立卷烟销售违规预测模型。使用BP神经网络算法对卷烟销售违规预测模型进行计算分析，经过多次训练不断调整参数最后找到误判率最低的预测模型，，利用该模型对卷烟销售的违规行为进行预测。 预测结果表明建立的模型预测精度较高，且具有很好的普适性，可以用来对各个零售户的违规情况进行有效预测，对可能发生违规的零售户进行预警。它不仅可以提高卷烟打假的准确度，还可以提高卷烟部门的工作效率，这为卷烟打假提供强有力的支撑以及帮助烟草专卖局制定相对适宜的卷烟抽查策略，从而实现精准打假。
[Abstract]:The tobacco industry is a high-profit industry, which is one of the main functions of the tobacco monopoly bureau. Because of the development of information technology, it is a problem to be discussed carefully at this stage. In this paper, the BP neural network technology is applied to the tobacco monopoly inspection, and it is of great theoretical and practical significance to explore the prediction model of the cigarette sales violation. Meaning. Through the qualitative analysis of the sales behavior and sales psychology of the cigarette retail, find out the related factors that affect the behavior of the retail account In ord to collect relevant data for a tobacco company in a city, pre-process that data, carry out quantitative analysis on a large amount of historical sales record and relevant data, and establish a cigarette sales violation prediction mode. the method comprises the following steps of: calculating and analyzing a cigarette sales violation prediction model by using a BP neural network algorithm, and finally finding a prediction model with the lowest judgment rate through a plurality of training and changing parameters, The prediction results show that the model forecasting accuracy is high, and it has good universality. It can be used to forecast the violation of each retail account effectively. The method not only can improve the accuracy of the cigarette making, but also can improve the working efficiency of the cigarette department,
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP183;F426.89
本文编号：2509016
[Abstract]:The tobacco industry is a high-profit industry, which is one of the main functions of the tobacco monopoly bureau. Because of the development of information technology, it is a problem to be discussed carefully at this stage. In this paper, the BP neural network technology is applied to the tobacco monopoly inspection, and it is of great theoretical and practical significance to explore the prediction model of the cigarette sales violation. Meaning. Through the qualitative analysis of the sales behavior and sales psychology of the cigarette retail, find out the related factors that affect the behavior of the retail account In ord to collect relevant data for a tobacco company in a city, pre-process that data, carry out quantitative analysis on a large amount of historical sales record and relevant data, and establish a cigarette sales violation prediction mode. the method comprises the following steps of: calculating and analyzing a cigarette sales violation prediction model by using a BP neural network algorithm, and finally finding a prediction model with the lowest judgment rate through a plurality of training and changing parameters, The prediction results show that the model forecasting accuracy is high, and it has good universality. It can be used to forecast the violation of each retail account effectively. The method not only can improve the accuracy of the cigarette making, but also can improve the working efficiency of the cigarette department,
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP183;F426.89
【参考文献】
相关期刊论文 前10条
1 李响;;基于BP神经网络的股价预测[J];大连海事大学学报;2008年S1期
2 司继文;韩莹莹;罗希;;Hedonic住宅特征价格模型的BP神经网络方法[J];管理学报;2012年07期
3 南凤兰;;证券公司风险预测研究[J];金融研究;2007年11期
4 刘洪,何光军;基于人工神经网络方法的上市公司经营失败预警研究[J];会计研究;2004年02期
5 史玉峰,曹俊如;基于遗传BP神经网络的变形数据分析处理[J];矿业研究与开发;2004年04期
6 岑成德,权净;服务属性对顾客满意感影响程度研究——人工神经网络方法[J];南开管理评论;2005年02期
7 刘永建;朱剑英;曾捷;;改进BP神经网络在发动机性能趋势分析和故障诊断中的应用[J];南京理工大学学报(自然科学版);2010年01期
8 卜耀华;;基于数据挖掘技术的企业财务困境预测建模[J];计算机仿真;2012年06期
9 石晓军;;Logistic违约率模型最优样本配比与分界点的模拟分析[J];数理统计与管理;2006年06期
10 李清;于萍;;财务危机预测主要方法比较研究[J];数理统计与管理;2012年04期
相关博士学位论文 前1条
1 姜明辉;商业银行个人信用评估组合预测方法研究[D];哈尔滨工业大学;2006年
相关硕士学位论文 前1条
1 陈鸿;人工神经网络在宏观经济预测中的应用研究[D];重庆大学;2005年
本文编号：2509016
本文链接：https://www.wllwen.com/guanlilunwen/shengchanguanlilunwen/2509016.html
