Financial Abnormal Data Detection System Based on Reinforcement Learning

Abstract

The fraud of financial data has seriously damaged the interests of stakeholders and hindered the healthy development of the economy. Based on the reinforcement learning theory, this paper proposes a model of intelligent analysis of enterprise financial data anomalies, constructs the difference method, and transforms nontemporal indicators into temporal indicators. The proposed model uses a convolution neural network with four hidden layers to classify, evaluate, and analyze the constructed temporal indicators of financial data. The main objective is a more effective and accurate identification of the issue. The test results show that the intelligent analysis method of financial abnormal data based on deep learning has ideal effectiveness and accuracy.