VR User Preference Recommendation Strategy Based on Improved BP Neural Network Algorithm
Min Wang, Fangwei Zhang · 2021 3rd International Conference on Artificial Intelligence and Advanced Manufacture · 2021
Digital twin technology and virtual augmented reality technology have been developed rapidly in recent years, and in order to make users get a good immersive experience, predicting user preferences and realizing integrated human-computer interaction are urgent problems. The application model of deep learning is a research hotspot in recent years, especially in some more complex pattern recognition problems, which has produced a large number of relevant research results, and the study of user preference strategy based on user line analysis, as a multidimensional data analysis task, is suitable for building multidimensional models to solve. Therefore, firstly, using the behavioral characteristics data of users in virtual reality, the Kmeans algorithm is used to categorize the scenes in virtual reality according to the obtained multidimensional data, and the data is enhanced to overcome the problem of unknown data characteristics and improve the robustness of the model. Again, according to the high-dimensional characteristics of users' behavioral characteristics, Improved BP neural network preference recommendation model was studied, and compared with the traditional clustering model for analysis. The experiments show that the model improves on the accuracy of user characteristics analysis in the virtual environment while ensuring the response time, and can effectively provide preference prediction to users, thus improving their immersive experience.