Research on Intelligent Technology for Economic Responsibility Audit Based on Text Classification and Semantic Recognition

In recent years, the rise of big data technology has provided a new approach for audit systems to utilize audit data to accelerate audit work, promote full audit coverage, and improve audit supervision levels. Text classification and semantic recognition technology are mainly applied in audit systems in areas such as text intelligent recognition, state perception, and reliability evaluation. They play a role in improving the comprehensiveness of audit coverage and the accuracy of audit data processing in conducting economic responsibility audits. In response to this, this article conducts research on the application of text classification and semantic recognition technology in the field of economic responsibility auditing, and proposes an intelligent audit data processing algorithm based on the BERT BiGRU algorithm. Firstly, the BERT model is used to vector represent the audit text of leaders fulfilling their economic responsibilities, and a vector containing semantic information is output; Then use the BiGRU algorithm to mine the dependency relationships between text vectors; Finally, the hidden state of the final calculation step of the BiGRU algorithm is input into the linear layer for classification, and the classification result is obtained through the layers. This article provides auxiliary decision-making for distinguishing responsibilities in economic responsibility auditing, thereby improving the accuracy and comprehensiveness of economic responsibility auditing work.