Forecast method of cigarette release based on hybrid neural network model
In order to scientifically formulate the cigarette product release strategy,a hybrid neural network model based cigarette release forecast method is proposed.By combining three methods:Boosting-GRU multivariate fore-cast model,product sales to release value calculation,and GRU multi-category model to achieve the generation of release strategies for different grades of cigarette products.The proposed method was validated by using historical sales data of cigarette products in Shiyan City,Hubei Province as the experimental object.The experimental results show that the method achieves an average accuracy of 97.64%in product sales volume forecast and the mean abso-lute error in release strategy generation were all below 20%.
万方数据
维普
