Research On Application Of Combination Forecasting Method In Cigarette Sales

As one of the pillar industries of my country's national economy,the tobacco industry plays an important role in safeguarding national and local fiscal revenues and social and economic development.In recent years,in order to promote the market development of the tobacco industry,the State Tobacco Monopoly Administration has successively proposed the"Key Points of the Reform of Cigarette Marketing Marketing Orientation".One of its core tasks is to explore scientific and accurate cigarette sales forecasting methods,improve forecasting accuracy,and strengthen cigarette sales.Application of predictions in tobacco companies.To this end,this article takes"cigarette sales forecast" as the end point,through the collation of relevant literature,summarizes the factors influencing cigarette sales and the method of forecasting cigarette sales,and takes Company Z as an example to do empirical analysis for the production of cigarettes by tobacco companies.Provide theoretical basis for task allocation,planning and decision-making.First,by combing,categorizing and refining relevant literature on cigarette sales forecasts at home and abroad,it is concluded that the factors affecting cigarette sales include political and legal factors,economic factors,and social factors;and take company Z as an example to make an empirical analysis.Select the quadratic curve regression method,the cubic exponential smoothing method,the gray prediction method,and the stepwise regression analysis method.Based on the cigarette sales data of Company Z from 2011 to 2020,the cigarette sales forecast models are established respectively,and then the percentage error method and the reciprocal mean square error are used.The method,entropy weight method,and Shapley value method combine each single prediction model to establish a combined model;then,the model evaluation index is introduced to compare the accuracy of each model,and the following conclusions are obtained:the combined model based on the percentage error and the single model and Compared with the combined model based on the reciprocal mean square error method,the entropy weight method,and the Shapley value method,the prediction value error evaluation indicators are all reduced,indicating that the method has the highest prediction accuracy,so the combined model based on the percentage error is finally selected and applied to Company Z Annual sales volume and cigarette sales forecasts of different price categories and specifications,after testing,it is judged that the predicted results of the model are in line with the expected sales volume.Finally,corresponding suggestions are put forward for tobacco companies from two aspects:the forecasting method of cigarette sales and the influencing factors of cigarette sales.