Research on Classification and Classification Model of Data Security Based on Graph Neural Network

Abstract

This paper proposes a data security classification and classification model based on graph neural network, aiming to realize the fine management of data assets through advanced neural network technology and expert knowledge. Firstly, a multi-level graph neural network structure is constructed, which can capture the complex dependency relationship between data and learn the deep features of data through the message passing mechanism between nodes. The model introduces an expert knowledge base to embed the experience and rules of domain experts into the learning process of the network, so as to enhance the interpretability and accuracy of the model. In this way, the model can not only automatically identify the security level of the data, but also classify the data in a fine granularity, thus providing strong support for the security management of the data. In order to verify the validity of the model, a series of simulation analyses are carried out in this paper. The experiment draws on real data sets from different industries, including finance, healthcare and education. The results show that compared with the traditional data classification methods, the accuracy and recall rate of this model are significantly improved. Especially when dealing with high dimensional and nonlinear data, the advantages of the model are more obvious. In addition, with the help of expert knowledge, the model can better adapt to the safety norms of specific industries, showing good generalization ability and practicability.