Audit Method Based on Event Extraction and Verification Warning for Power Engineering Projects

Due to the wide scope of business, multiple audit processes, and complex event correlation in power engineering, it is difficult to achieve digital operations in the audit business. This article proposes a power scalable business audit method. Firstly, by constructing an audit business process based on an event framework, the workload and business expansion pressure of domain experts can be reduced. Subsequently, the Transformer combined with template matching was used for event extraction. Use the identified audit events of fixed documents as remote knowledge to generate trigger word templates with clear prompt knowledge. This method can supplement remote knowledge as needed based on different project themes, thereby improving the accuracy of event extraction. Finally, the identified audit events are assigned to different audit stages and event frameworks. This is used to support the verification and warning of related audit events during the process. Experiments have shown that in low manual labeled text, the accuracy of event extraction is higher than 93% by this method. Extracting events can accurately match various audit business processes, which supports verification warning of related events.