STUDY ON EARLY WARNING OF ENTERPRISE FINANCIAL DISTRESS - BASED ON PARTIAL LEAST-SQUARES LOGISTIC REGRESSION

Establishment of an effective early warning system can make the company operators make relevant decisions as soon as possible when finding the crisis, improve the operating results and financial condition of enterprise, and can also make investors avoid or reduce investment losses. This paper applies the partial least-squares logistic regression model for the analysis on early warning of enterprise financial distress in consideration of quite sensitive characteristics of common logistic model for the multicollinearity. The data of real estate industry listed companies in China are used to compare and analyze the early warning of financial distress by using the logistic model and the partial least- squares logistic model, respectively. The study results show that compared with the common logistic regression model, the applicability of partial least- squares logistic model is stronger due to its eliminating multicollinearity problem among various early warning indicators.