Research on Auditing Supply Chain Finance Business of State-owned Enterprises Based on Deep Learning

Abstract

Due to the complexity and tediousness of the current audit process, the development of audit intelligence has become a general trend. In order to improve the audit quality, the study establishes an intelligent financial audit model based on audit opinion for the supply chain finance business of state-owned enterprises after analyzing the application of the audit function of deep learning, for which an audit prediction model based on the Gray Wolf Optimization Algorithm (GWO-Optimization) and the fusion of Long and Short Term Memory Network (LSTMN) is proposed. The supplier of a state-owned enterprise is selected as the research object, and the GWO-LSTM model is trained and tested by constructing the audit opinion prediction index system and data collection, comparing it with the BP neural network and support vector machine model, and combining it with the gray prediction model for predicting the audit opinion of the samples, in order to improve the model’s practical application ability. The GWO-LSTM model performs better in predicting audit opinions than the comparison models, as evidenced by the results. Its prediction and training accuracies are above 80%, and the accuracies of RMSE, MAE, and R² are finally stabilized at 0.1, 0.1, and 0.948, and the combined accuracy of prediction in practical application reaches 90%. The model in this paper can scientifically predict the audit opinion, thus improving the efficiency of the audit data analysis for the supply chain finance business of state-owned enterprises.

Keywords

• gray wolf optimization algorithm
• long and short-term memory network
• audit prediction model
• gray prediction model
• deep learning
• supply chain finance
• 03b70