Research On The Application Of Combination Forecast In Medicine Circulation Enterprises’ Sales Forecast

The rule of drug sales is not a simple linear behavior, but a nonlinear behavior affected by factors such as market, economy and policy. It contains uncertainty, ambiguity and inevitability. Meanwhile, it contains some kind of inevitability. With limited amount of sales data, choosing reasonable methods for medicine circulation enterprises to carry out reasonable invoicing and reduce operating costs has huge guiding significance.Firstly, this paper introduced the necessity of accurate sales forecasting, reviewed the research status inside and outside on issues related and made a summary of main forecasting methods. Secondly, for problems in the drug sales process, such as linearity, nonlinearity and uncertainty, the researcher selected three forecasting models, introduced relevant theories and summarized advantages and disadvantages of them. Thirdly, the researcher proposed a variable weight restructuring forecasting model called CAMB which was a non-linear combination of neural network forecasting model. Meanwhile the researcher gave the algorithm processes and theoretically elaborated the superiority of combination forecasting model with respect to the single forecasting model. Finally, this paper applied sales data to each model and verified the effectiveness and practicability of CAMB model through analyzing the results.It’s impossible to ensure that the forecasting model can get better results in any cases. CAMB combination forecasting model presented in this paper consisted of time series model, cloud model and neural network model. It made full use of the advantages of each model and synthesized the different aspects of useful information to get optimum combination forecasting model. It reduced the risk of a single model for sales forecasting and provided a new way to improve the prediction accuracy.