E-commerce consumer behaviour prediction through the integration of collaborative filtering and graph neural networks

Abstract

Due to the huge amount of information interested by users on e-commerce platforms, it is difficult to predict consumers’ purchasing behaviour. To this end, this paper first forms a session graph based on consumers’ session sequences. Meanwhile, inter-item multivariate relationships and inter-session cross-information are modelled through graph convolutional networks. Then, the user’s intention representation is generated through comparative learning. Next, a behavioural model of user consumption based on attention network is constructed. Finally, this paper calculated the ratings of users with purchasing behaviours on the target items, and obtained several items with high ratings to generate a recommendation list to predict e-commerce consumers’ behaviours. Experiments are conducted on two public datasets, and the results show that the accuracy of the proposed model is improved by at least 3.31% and 5.21% respectively, which effectively improves the accuracy of e-commerce consumer behaviour prediction.