Prediction of consumer repurchase behavior based on LSTM neural network model

To clarify the factors that affect consumer desire to purchase and promote the development of the market economy, the repurchase behavior of e-commerce platform users is used as the background to study how to use edge computing to collect customer shopping data accurately. Consumer shopping behavior is predicted by a consumer shopping information data platform built with edge computing technology, and is modeled by a joint model of Long-Short Term Memory neural network model and convolutional neural network model. The prediction accuracy of the neural network model is verified through the analysis of the prediction results, and on this basis, a method of information segmentation processing is proposed to further improve the prediction accuracy of the neural network model for consumer shopping behavior. The results show that information segmentation processing can improve the prediction accuracy of a variety of neural network models by more than 2%, and even increase the prediction accuracy of neural network models based on Extreme Gradient Boosting by 5.4%. From this point of view, it is feasible to use digital technology to predict consumer repurchase behavior, and mathematical modeling based on various neural networks plays an important role in the study of consumer repurchase behavior.