With the arrival of the digital era, the scale of data is getting bigger and bigger, and the traditional auditing methods, including sampling audit, circular audit and other methods, can no longer meet the current needs. In this regard, this paper uses association rule algorithm to complete the data mining work on the basis of laws and regulations database, financial and business database, obtains a total of 30 financial and business association rules, and verifies their usability by calculating evaluation indexes. From the definition of anomaly rules and the principle of outlier detection, the anomaly detection model is designed to facilitate the understanding, judgment and audit of business personnel. Finally, under the theoretical guidance of EA-LDA combination algorithm, the hidden relationships between financial statement auditing entities are obtained, which are added to the original database, and finally the financial statement auditing knowledge graph is constructed and analyzed by in-depth exploration. After analyzing, the difference degree of financial statement auditing entities is less than 0.1, and the probability value is less than 0.5, and at the same time, some examples of financial statement auditing knowledge graph based on the EA-LDA combination algorithm are shown, which fully verifies the practical research value of this paper. This paper has certain reference value for enterprise financial statement auditing, which in turn improves its audit quality and efficiency, with a view to ensuring the reliability of its financial statements.
