Financial fraud risk analysis based on audit information knowledge graph

The frequent occurrence of financial fraud caused by listed companies has seriously hindered the healthy development of the capital market. Existing studies usually analyze financial fraud through the effectiveness of audit opinion or the correlation between auditor change rate and financial fraud, ignoring the relationship between different participants. By using the audit relationships among corporations, audit firms and auditors, this paper constructs an audit information knowledge graph and proposes a knowledge graph reasoning framework based on the Sub Feature Extraction method to detect potential fraud corporations. In the process of analyzing the audit data of 376 companies in the China Growth Enterprises Market from 2013 to 2019, it is found that potential fraud corporations can be well identified by searching from the known fraud corporations using searched paths. In addition, we find two new audit features of financial fraud corporations which are respectively related to abnormal audit opinions issued by auditors and abnormal associations of audit firms. They can help regulators more effectively find the potential financial fraud corporations that need to be supervised. (C) 2021 The Authors. Published by Elsevier B.V.