Research on User Portrait Technology Based on Dynamic Graph Neural Network

In recent years, with the popularity of smart wearable devices, fine portraits of human behavior based on angular velocity, acceleration and other sensors have become a research field of concern. The traditional human body fine portrait method relies on manual feature extraction and rule-based method, which has the problems of low recognition accuracy and poor generalization ability. In this paper, a human behavior portrait method based on dynamic graph neural network is proposed, which realizes the efficient recognition of complex behavior by constructing dynamic graph structure and extracting features by graph convolution operation. The experimental results show that the proposed method has superior performance in human behavior recognition tasks and has a wide range of applications.