Research on 3D advertising placement based on virtual reality simulation

Abstract

Virtual reality based on computer 3D technology has brought revolutionary changes in advertising creativity. To enhance the economic value of advertising itself, it is necessary to provide personalized and precise advertising placement for users from a massive advertising database. A recommendation system based on deep graph convolutional networks is a personalized system that can recommend based on their attributes and historical records. However, complex models are required to learn higher-order feature information, and the mining of user interaction behavior is insufficient. This article proposes a graph convolutional network with multi-head attention embedding for 3D advertising placement. We construct advertising graph data based on user behavior and multi-feature embedding mapping. To exploit higher-order features, a multi-head attention mechanism is embedded in the graph convolutional network. The experimental results show that the proposed model achieves better performance compared to other depth models on Criteo and Avazu datasets. We will deploy the entire model on the server to achieve precise advertising placement.