Online Product Sales Prediction Based on Elman Neural Network

The prediction of online product sales affects the decision-making process of enterprise online sales. The academic community has conducted numerous studies on product sales prediction, and the vast majority of existing literature uses feedforward neural networks for sales prediction. As sales belong to time series data, recurrent neural networks with memory function may be more suitable for online product sales prediction. Therefore, this paper constructs the Elman neural network model as a representative of recurrent neural networks, and constructs the BP (Back Propagation) neural network model as a representative of feedforward neural networks. By using the same data to train and test the Elman neural network and the BP neural network, the error between the two network networks is compared and analyzed. The results show that the Elman neural network's prediction of online product sales is closer to the actual value. The mean absolute error (MAE), root mean square error (RMSE) and mean absolute percentage error (MAPE) of the BP neural network are 8.03 times, 8.12 times and 8.24 times of the corresponding values of the Elman neural network, respectively. The error value of the BP neural network is much larger than that of the Elman neural network, and the Elman neural network has more stable and less fluctuating prediction. Therefore, compared with the BP neural network, the Elman neural network has higher prediction accuracy, more stable prediction, better prediction effect and more accurate prediction of online product sales.