Font Size: a A A

Research On Risk Of Listed Companies In China Based On Data Mining Technology

Posted on:2019-03-09Degree:MasterType:Thesis
Country:ChinaCandidate:Y YanFull Text:PDF
GTID:2359330545490147Subject:Statistics
Abstract/Summary:PDF Full Text Request
With the continuous increase of domestic economic downward pressure and deepening of economic transformation and upgrading,the operating pressure of individual listed companies is increasing,and the risk of listed companies' delisting is also gradually emerging.Once a listed company is crowned with ST and*ST hats,it will have greater pressure.The financing difficulty of ST will further increase and its operation will deteriorate is due to the strict supervision of relevant departments.Even if enterprises still losing money for third years,they will be suspended from the market.We select all the listed companies that are first specially dealt with by the CSRC and all the delisting enterprises as risk samples in A shares from 2012 to 2017,with a total of 137 samples.After that,Combined with the F fraction model,137 of the normal enterprises and the unstable enterprises are matched according to the same industry,the same period and the same asset scale.The selected indicators cover the financial and non-financial information of the listed companies.Then,we get 9 indicators through the index screening to distinguish the risk of the enterprise.We build enterprise risk prediction model by the data of t-3 year enterprise sample through the Logistic regression,neural network,decision tree and random forest algorithms.From the comparison of prediction models,the stochastic forest algorithm is superior to the decision tree model,Logistic regression and neural network model in predicting the risk data of listed companies.The overall discrimination accuracy of the training set is 99%,and the overall accuracy rate of the test set is 62.2%.And through the rules produced by the decision tree,it is found that whether the interest guarantee multiplier is greater than 6.7 is a distinction between the normal enterprise category and the other categories.From the stochastic forest model,we also found that the interest guarantee multiple and the operating profit ratio are more obvious to the classification of the enterprise category than the other indicators.The risk problem of listed companies will not only bring huge losses to enterprises,but also damage the interests of investors,creditors and other stakeholders.It seriously threaten the survival and development of enterprises.How to find out the risk information from the numerous data of the enterprise and prevent the risk ahead of time.It have the great significance for the healthy development of the enterprise.
Keywords/Search Tags:listed companies, F score, risk prediction, decision tree, random forest
PDF Full Text Request
Related items