Research on Intelligent Identification Model of Financial Fraud of Listed Company Based on Transformer Architecture

At a time when the world is overwhelmed by the wave of digital transformation, the methods of financial fraud applied by companies on the stock market have undergone unprecedented technological change. Fraud techniques have changed from traditional accounting manipulations to" intelligent " scams that use advanced technologies such as large-scale data analysis and artificial intelligence. It has new features such as robust protection, high technical level and complex evidence chains. Based on an analysis of China's capital market, this study focuses on three main difficulties in identifying financial fraud: enormous difficulties in processing data, complex obstacles to determining correlation, and the need for real-time monitoring. Financial fraud can be achieved by relying on the transformer architecture as the underlying technology and by forming a three-dimensional data collection network and a dynamic risk assessment model. This study, which enables the identification of behavior with great precision, provides a possible technical approach to current financial fraud management and sets the direction for the development of financial supervision reform in the era of the digital economy.