Financial Failure Warning Of China Listed Companies

Increasingly large scale listed company group play an important role in promoting our country to establish and perfect modern enterprise system, realize resource allocation optimization and deepen the reform of economic system, and its management quality has impact on China securities market directly. The listed company facing more and more competition and market uncertainty, thus more likely happened financial failure. Listed companies'failure has pernicious influence in different interest parties and the whole investment market. Therefore, the warning research of listed companies'failure has great significance.Throughout the current enterprise financial failure warning research, mostly using parameters discriminate analysis method, and nonparametric density estimation have good properties, but rarely apply in actual particularly economic fields. This paper will apply k-nearest-neighbors of nonparametric density method to listed companies'failure discrimination.This paper define a listed company as a failed company if it has been disposed by securities regulators with ST, select 251 failed listed companies and 251 matched healthy listed companies as research sample, design a series of index variables, and use Logit method, Discriminant Analysis and K-nearest neighbors method separately to construct models.Through empirical study, it turns out that k-nearest-neighbors model performs close to Logit model and better than Discriminant Analysis model, and have a good promotion application ability. At the same time, we found that industry factors have significant effect on early warning research of enterprise failure. And establish an index system for the relative research reference.