Establishment of a Combined Sales Prediction Model Based on BP Neural Network

Enterprise needs more accurate sales data to make plan for the future as the global economy is falling into a slow growth period because of the blocked logistics caused by worldwide spreading covid-19, but traditional single prediction model such as triple order exponential smoothing prediction model, Grey prediction model and LSTM prediction model is difficult to get accurate prediction output due to their limitation. This paper proposes a BP neural network combined prediction model based on multiple models, which can get accurate output by inputting results of mentioned models above. It has applied to medical sales industry and is recognized for meeting the desired goals.