Research On Forcast Model Of Cigarettes Sales Quantity

China has accessed to the WTO. With tariffs on foreign cigarettess droping and other relevant provisions of opening up, tobacco industry gradually realized that Chinese tobacco industry will face fierce competition. So the research on prediction of cigarettess sales quantity is very important to enhancing the competitiveness of Chinese tobacco industry and has important application value. According to the characteristics of Chinaese tobacco industry, this paper chooses the artificial neural network method to establish the forcast model of cigarettes sales quantity.In the first, the paper summarizes the characteristics of Chinese tobacco monopoly system and tobacco consumption and analyzes the main factors which influence the demand and sales quantity of cigarettes.Secondly, the paper chooses the calculation of forecast indexes according to the factors that affect cigarettes sales quantity, reduces the dimensions of indexes by using clustering method , then cuts out the redundant indicators by using gray relational analysis method and establishes the forcast model of cigarettes sales quantity based on artificial neural network. At the last, this paper applies this model to forcast the total monthly sales quantity of an area of Anhui province. Results show that the forcasting effect of the monthes in non-New Year period is good but not good in New Year period. According to the characteristics of cigarettes sales quantity data, the paper establishes the gray model by using subtraction idea to forecast them. The results are good. It can be seen that, the prediction effect on monthly sales quantity in non-New Year monthes is good when using prediction model of the cigarettes sales quantity based on artificial neural network. When we forcast the monthly sales quantity in New Year period, we shoule use the gray model based on subtraction idea.