| China is the world's largest country in the production and consumption of tobacco products,which accounts for about 40% of the world's total.The consumer groups have characteristics such as a large population,a hierarchical age structure,and a diversified consumer demand.In recent years,in order to adapt to the complex and ever-changing market environment and create an efficient and agile supply chain system,China's tobacco industry has undergone a marketoriented reform.One of the core contents is to explore a more scientific forecasting system for the market and improve the quality of forecasting.Based on the background,this paper try to focus on these fields.Early studies on cigarette demand forecasting exist some shortcomings such as the market forecasting factors of tobacco products in the tobacco industry are relatively unique and can neither effectively reflect the impact of different types of factors on the forecast results,nor overall forecast with different statistical period data.In addition,traditional forecasting methods involve time series processing and regression analysis and others,due to their limitations,these methods have some problems in the construction of prediction models and the effect of actual prediction.In response to above problems,this paper analyzes the factors that affect the market forecast of cigarette products and classifies them into three categories: Plan factors,Market factors and Condition factors.Then artificial neural network is introduced to compare the forecasting results of different methods through empirical analysis.The results show that artificial neural network prediction can well meet the requirements of diversified and nonlinear factors in the dynamic environment during the forecasting process.It not only overcomes the shortcomings of existing predictions,enhance the quality of prediction,but also comprehensively analyzes different kinds of factors to further improve the accuracy and reliability of forecast results. |
PDF Full Text Request