GAN-Based Financial Data Generation and Prediction: Improving The Authenticity and Prediction Ability of Financial Statements

Abstract

The research on the mining algorithm of financial data association relationship mainly explores a certain kind of association relationship in depth, but it is not suitable for the attributes and characteristics of financial data itself, and there are few comprehensive analysis and application for financial data association relationship mining. In order to overcome the above problems, this paper proposes a financial data generation and prediction model based on GAN. Based on WGAN network, this paper improves the authenticity of the generated virtual samples by increasing the cyclic consistency loss term and selecting intermediate samples for the generated samples to optimize the generated model. At the same time, in the system, this paper adopts intelligent data analysis research method, mines the correlation of different dimensions of financial statement data, and presents the mining results by using the correlation visualization method, so as to realize the risk assessment and trend prediction of enterprise financial status. According to the comprehensive experimental analysis results, it can be seen that the model proposed in this paper has good performance in the authenticity analysis and prediction of financial data. Generally speaking, the model proposed in this paper provides a reliable tool for the authenticity audit of financial data, and can provide a reference for the formulation of subsequent schemes and policies through financial data prediction.

翻译

搜索

复制