Based On Time Sequence And Multiple Linear Regression Model Of Rural Cigarette Sales Forecasting

The tobacco industry as one of the important industry of China, play crucial role to the national economic development and fiscal revenue. At present, the tobacco industry is facing a realistic predicament that the foreign cigarettes make inroads on our market shares gradually. By reason of the institution that tobacco industry under the national control distributes cigarettes according to indent’s quantity, it will forecast the cigarette sales. The purpose is that improve the competitiveness of China’s tobacco industry more effectively, reduce the inventory cost of the enterprise and arrange production plan more reasonable. The research object is cigarette sales in the countryside of Yunnan province, owing to cigarette’s regional problem and the importance of the rural market.First of all, this paper summarizes the different methods in the application of sales forecasting and the current research status of factors affecting cigarette sales. In the next place, it determines the research route and prediction index of cigarette sales. It is a next step that process the data which can’t be applied directly. It analyzes the processed date by the method of time series and multiple linear regression.Moreover,it takes appropriate method to establish model of time series and multivariate linear regression. It makes the comprehensive model using predicted value of the above models.Comprehensive model not only considers the influence of the early stage of the internal data, but it also takes the influence of the external macro factors into account, overcoming the problems that the neural network model can’t output the size of the influence of independent variable on the dependent variable. In the theoretical model, the influence of independent variable on the dependent variable is verified that is about the size effect of associated with the value of the early stage of the dependent variable. Comprehensive model is better than the time series model or the multiple linear regression model in Yunnan province’s cigarette sales forecasting.Thus,the mathematical models of comprehensive has good meaning in theory and in fact.