Research On Financial Risk Evaluation Of Digital China Information System Co.,LTD Company On Deep Neural Network

Financial risk is one of the main risks to be guarded against in the process of enterprise development.Financial risk widely exists in the process of enterprise operation or decision-making process,threatening the healthy operation of enterprises.It is very important for the safe and stable operation of enterprises to pay attention to the financial situation of enterprises and to evaluate and warn the possible financial risks of enterprises.Academic evaluation of enterprise financial risk has always been one of the important subject directions,through methods and tools to evaluate the financial situation of enterprises,the possible risk prevention and early warning,will become an important academic research results.This thesis adopts the model evaluation method,uses TensorFlow framework and DNN deep neural network model,combines the typical operational and financial characteristics of IT service industry,and takes the solvency,operational,profitability and development capabilities of enterprises as the risk evaluation criteria.The financial data of 63 companies in recent four years are collected.Established a neural network can be used in the IT services financial risk evaluation model,the typical representative of IT services-China information,the evaluation of financial risk according to the evaluation results,compared the enterprise the typical competition in industry at present,has carried on the further analysis and evaluation,the objective evaluation results.In the study of this thesis,combined with the previous research experience to carry out the following extended work.Firstly,in the process of judging financial risk labels,the SPSSPRO tool was used to judge the risk of 252 financial statements of listed companies.The missing data filling,Z-score data standardization,data dimensionality reduction,PCA principal component analysis,and cluster analysis were respectively used to obtain the labels of 252 financial statements with and without risk.It provides the data basis for the subsequent modeling.Secondly,a neural network model was established by using TensorFlow framework and deep neural network method.252 data were studied,tested and verified,and the data accuracy was relatively high.Finally,after the evaluation results of deep neural network are obtained,the comparative research method is used to conduct a peer comparison analysis of Shenzhou Information Corporation to verify the evaluation results of deep neural network and obtain relatively objective evaluation conclusions.Finally,combining a variety of research methods in this thesis,based on the depth of the neural network model,combining with the evaluation results,the shenzhou information is given suggestion to certain risks,avoid future financial risks for the companies to provide a certain train of thought,enrich the existing literatures and research methods,providing a certain reference to the research of the subsequent scholars.