Research on Adaptive Repair Algorithm for Intelligent Processing of Computerized Data in ERP System Using Graph Neural Network

This paper studies the adaptive repair algorithm of graph neural network (GNN) in computerized data processing of ERP system to achieve efficient and intelligent data processing. This paper proposes an adaptive repair algorithm based on graph neural network, which models the computerized data in the ERP system as a graph structure, and performs deep feature extraction and anomaly detection on the data through the relationship between nodes and edges. A multi-layer graph convolutional network is used to realize feature aggregation and nonlinear transformation, and the model parameters are optimized through adaptive learning strategy, so that it can dynamically adjust the processing process and enhance the accuracy and flexibility of data repair. To verify the effectiveness of the algorithm, this paper conducts multiple model simulations on data sets of different sizes. The simulation results show that this method is significantly superior to traditional algorithms in repair accuracy and processing efficiency, especially in high-noise data environments. It still shows strong robustness. Data analysis shows that the repair accuracy of the model in different ERP data scenarios is improved by an average of 18.5%, and the processing speed is accelerated by 22.3%, which proves the potential of combining graph neural network and adaptive repair algorithm in intelligent data processing of ERP system.