Research On Financial Distress Identification Model Of Listed Companies Based On Improved Deep LSTM

By the end of 2022,there were more than 5,000 A-share listed companies on the Shanghai and Shenzhen stock exchanges,but the law of survival of the fittest does not stop companies from companies.internal management and unpredictable competitive environment have increased the operational risks of listed companies in China.In order to alert investors and regulators to avoid risks in advance,it is imperative to establish a set of effective financial distress identification model.In this paper,after reviewing the literature research on the financial distress identification model,firstly,the definition of financial distress is defined,and the root causes of financial distress and various internal and external factors are deeply analyzed.Secondly,on the basis of the existing financial distress identification model index system,risk factors such as audit opinions and macro environmental factors are introduced.and a set of financial distress identification model index system consisting of 25indicators such as current ratio and cash ratio is constructed this paper chooses the LSTM neural network model for empirical study.By building a four-layer neural network model that includes Dropout layer and BN layer,and module to introduce the attention mechanism,based on the data of A-share listed companies in Shanghai and Shenzhen from 2018 to 2021,the financial distress identification model is trained from the single-step dimension,and the prediction accuracy of the artificial neural network model under different time dimensions is obtained.The research findings in this paper are as follows:The application of the optimized index system and the improved LSTM neural network model to the construction of financial distress model can make the prediction results have better accuracy and stability.The prediction accuracy of this model in T-1,T-2 and T-3 years is 94.54%,93.89%and 93.45%,respectively.Investors and regulators can substitute the relevant data of the target company the target company is in financial distress.Finally,this paper constructs a financial distress identification model suitable for Chinese listed companies,so as to reveal the listed companies with financial problems to investors and regulators,so as to arouse investors’ alarm and reduce risks.At the same time,it can also help regulators maintain the stability of the capital market order and provide timely scientific enterprise information for the market.