A PREDICTION MODEL OF USER BUYING BEHAVIOR BASED ON LSTM AND SVM
Xue Bingbing, Jitao Zhao, Wei Xianyi · 2021
The rise of e-commerce has accumulated a large number of online consumption behavior data of users. Analysis of users' purchase behavior provides support for users' decision-making or merchants' marketing. A method for predicting user purchasing behavior based on Long short-term memory and support vector machines is proposed by this paper. First, it uses LSTM to automatically extract and select characteristics, saving the manpower and time for extracting features. At the same time, it extracts dynamic characteristics of user behavior data as successively related sequences as the advantages of LSTM on sequence data. Then, we predict user purchase behavior by support vector machine. Finally, we use the real monthly sales data of JD for experimental verification. The experimental results show that the accuracy of the model is higher than other models. So the model has a great application prospect in the prediction of users' purchase behavior.