Research On Financial Distress Prediction Of Information Technology Listed Companies Based On LightGBM

With the coming of the big data era in recent years,the concept of 5G,Internet plus,big data,cloud computing and block chain has been constantly put forward,especially the rapid rise of Internet Co,which brings full economic vitality to the country.Under the policy support of the 13 th five year plan,information technology industry has gradually become a pillar industry of national economic growth,an important driver of high-quality and rapid economic development in China,and also a leading and strategic industry leading national production and life.With the rapid development of information technology industry,its characteristics of high growth and high risk are more and more prominent.Due to the problems of high technical iteration requirements,weak solvency and unstable cash flow,the industry is more likely to have potential financial risks and even financial crisis.In 2019,there are 16 information technology enterprises that were st,accounting for 17.9% of the total st enterprises in the whole year,only behind the manufacturing industry.Therefore,the information technology industry urgently needs a targeted financial crisis prediction model to help enterprises predict whether there are serious financial risks in advance.Based on the research at home and abroad,this paper takes 102 listed companies of information technology industry in Shanghai and Shenzhen A-share markets in 2010-2019 as the research object,selects the financial data and non-financial data of T-3,and selects the stock price information of 500 trading days before T-3,builds the network by using the complex network theory,extracts the corresponding network indicators such as centrality,Page Rank value,etc.,and combines them The financial indicators and non-financial indicators of the enterprise are 59 indicators in total,and a comprehensive model index system is constructed.Aiming at the excellent performance of lightGBM algorithm,this paper uses ligh GBM integrated algorithm to build the financial crisis prediction model of information technology listed companies,and adds the network index information.After empirical research and analysis,the accuracy of classification reaches 92.06%,and the AUC value of the model reaches 0.968.Compared with LR,SVM,RF and other different models,the model shows a good prediction effect,which has a good guiding significance for information and industry crisis prediction.At the same time,this paper makes a comparative analysis of the model whether it contains financial network indicators.The results show that network indicators play a positive role in judging whether an enterprise has a financial crisis,and enterprises with high importance in the network are less likely to have a financial crisis.